Recently laid off Amazon Robotics Engineer, I think your skepticism is spot on, some of the skillsets of a developer can translate well. The problem is these big companies often paint a grandiose picture of their robotics for press release and rarely deliver, usually do to economic or Hardware issues. Hardware trails software, and you have to take a hardware first approach as Steve jobs mentioned. In addition, companies like amazon will have 2-3 separate projects that essentially do the same thing and compete internally so they can hedge their bets. The engineers on the "losing" teams often end up looking for work elsewhere, just look at projects like Scout.
Correct. Right now the center of hw & electronics in the US is SXSW & Mexico. But it still trails significantly to South East Asia. But the Americas is starting to bring that back. It will be exciting to see what it looks like in 10 years.
Firing engineers for non-engoneering mistakes is stupid and wasteful. 2-3 teams competing on different systems design is fine but that's a special role who makes those architectural decisions. Software/mechanical engineers execute those designs and must be fired only if they fail in their own engineering crafts.
I've been watching a lot of low-level programming channels lately and started dabbling with C. Systems engineering is a field that I would like to get into soon... sounds way cooler than building the 100th version of the same crud app.
Yeah, no kidding. And the crud app is exactly the bread and butter for the low/no code world. Honestly I really don't like low/no code as a dev all that much. But I finally sat down and learned bubble and 1) it certainly can build a LOT of stuff 2) a lot of what it can't do can be added via just a few custom REST services or whatever. Can't do everything but it sure can do a lot.
I'm a robotics and AI engineer, with a background in Mechanical engineering and went to grad school for robotics and AI. It's not an easy pivot or a career to go and chill 😅. Robotics of previous decades is not the same as modern robotics. Before, you needed to understand electromechanical fundamentals and you could work your way through the rest. In today's robotics, you need to understand some new fundamentals you need to have development skills to start with then add on the fundamentals of AI and background knowledge in computer science. Then you need to choose your focus whether it's manipulators, Mobility locomotion, Sensing and perception, Operator interface, Manipulators, and effectors each of those requires programming skills. The same computer science fundamentals are required when building autonomous systems (could be vehicles or some other system). It's also not that easy for the companies building these systems to make them work and produce them in mass at a profitable cost and retain their staff. It's a long-term investment career and I would consider it as R&D. You can also go into automation & manufacturing which is more accessible and stable as a career since manufacturing plants use automation anyway.
Thanks for the details! Would you be open/interested in doing an online interview about what it takes to get into the field and maybe your thoughts on the big $ investments in the news? If so drop me your contact info changenode.com/contact/ :)
thank you for your comment! Im a student doing my Master in robotics and autonomous systems. it was pretty tough for me and i still struggle to image where I can work in the future. How to find out?
@@xptransformation3564 your first step IMHO is to start calling/networking with recruiters in your area and see what they have to say. Your education is an excellent first step - you need to fold in talking with hiring managers ASAP. Good luck!
20 years in IT here, expertise in development, masters in computer science & AI degree. On each AI job there are 100 applicants within an hour. Recruiter told me 400+ Media make rumors that does not seem to have translation in reality. btw I was laid off because it was no AI work any more. That's why I doubt all these 'exciting' news are reliable.
@@rafae5902 Depends on the job, eg I know someone who works on Android Automotive and someone else who does IoT stuff all from home, but for a bunch of it I would think you'd have to go in. The devices for both of those guys are pretty small, so no biggie to set up a dev desk at home. I could see where for some of the other stuff it might be work hybrid, but still have to go in at least 1-2 days a week.
Robotics engineer here - labour cost alone is not a meaningful metric. Human presence reduces opportunities to automate (a lot of safety-related limitations). Even if cost is only 10% now, removing a human altogether can allow you to build cars a few times faster, so cost savings will be way above 10% that you'll shave off human labour. And don't hold your breath on robots co-existing with humans. Even if technology will deliver (spoiler alert - it won't), the regulations will catch up years later. Large red e-stop buttons will be pervasive for years to come. As for 3D-printing - not that useless, 3D-printing unlocked generative design. You cannot machine such shapes with the regular subtractive machining, but metal 3D-printing can do all those organic shapes easily. In prototypes or in scale - does not matter.
Yeah, tl;dr I agree with all of this over the long term. The Volvo plant is basically new, so that's why I used it at an example of what exists today. Any thought on timeline for that kind of factory going dark? RE 3d printing in particular there's a bit of a terminology issue. I'm kind of short handing 3d printing as the home units. Technically I'm sort of confusing/labeling an industrial scale 3d printer as industrial automation. The point I was going for is that we don't really do 3d printing at scale at home. Thanks!
Maybe the real problem to solve is how do you introduce robots with humans present and still get the benefits without sacrificing safety. The Army is funding research for drone medevacs, where an unmanned helicopter carries a patient from a remote location. The solution isn't to get rid of humans so that robots can do their jobs more efficiently.
@@cryora and this is an incredibly hard problem to solve. Soft robotics (e.g., pneumatic actuators) can be promising, but there is a lot of unsolved issues with it. The usual hydraulic or direct drive electric actuation is not yet suitable for operating in the same space as humans - even if sensing and control is solved, there is still a need to prove that it is 100% perfect and glitches won't happen. Regulators at the moment only accept the full shutdown, no actuators powered as a safety threshold for humans (e.g., in industrial robotic cells).
I've seen software demos that were literally nothing but made up screenshots/animations purported to be shipping software. What's your take on the figure.ai demo & timeline...?
@@ChangeNode Very hard to say, but I would say their biggest hurdle will be regulation as their bot is technically a cobot. I'm my opinion configurable Microfactories are a more realistic medium-term solution.
As someone who wanted to get into robotics from the start but no one considered me because I didn't have an engineering degree, I see this as an absolute win. But I am not gonna apply to the Amazons and what not. I am thinking more the small start-ups.
I didn't knew about this channel. Instant favorite. Thank you for this awesome content (I'm a Bsc in EE, transitioning to Machine Learning and Robotics through masters and freelancing).
Truer words have never been spoken 🎉 I have been looking at Zig as a replacement for C/C++ and has a much lower learning curve compared to Rust, at least for me. Looking forward to how the robotics industry shapes up
Interesting. Just took at look at blog.logrocket.com/comparing-rust-vs-zig-performance-safety-more/ and the comments about FFI caught my eye. I've heard more about Rust in the context of the Linux kernel, and I've played with it a bit in the context of Tauri. Haven't seen Zig come up.
Zig is cool, but I have been using Free Pascal and love the way it forces stronger code typing, Ada would be even better. Not sure why Ada is not used more.
This is the first video of your channel that I've seen after watching enough other types of tech commentary that 90% of the time just revolves around which language is better, which editor is better, this tired old measurement of tools instead of products, achievements and good investment of time, I've almost slam shut my tablet after hearing how 80 hours per week is the way to go while 50 is enough to maintain sanity. So this mature analysis of the state of affairs, and proper weighing more real-world aspects of the economics behind re-training the workforce, the natural resource dictatorships bent on violent expansion occasionally, which I've mentioned in discussions myself is so refreshing. Thank you for this video!
On manufacturing automation: 4:55 I think the 5-10% labour cost is likely for current huge operations benefitting from mass production/economies of scale. Labour cost as a percentage is far higher on smaller, more bespoke production, so that’s where I see the more impact coming eventually, as automation costs drop. Cheaper, more capable automation is likely to mean more flexible ‘ cheaper operation on smaller scales, allowing more customised/variable production, faster turnaround & lead times, and shorter production runs becoming more feasible. That could open the door for different kinds of [re]manufacturing on a local scale.
Thanks, this video made me think. I've been an enterprise developer for almost 15 years using mostly scripting language. Last year I took a 10-month C++ course. It was quite an awesome experience to manage memory yourself. Hoping to find a way to switch my career to something more challenging like embedded, or just working on more profound software products for desktop/servers, like OSs, browsers, game engines. Robotics field looks intriguing, but I'm afraid I lack electronics/math skills to get such kind of jobs.
Robotics is a multi-disciplinary field, you need knowledge of embedded systems, low level programming, AI/ML, control theory, path planning and estimation, mechatronics and a lot of maths. You can specialize in one of these domains and still get work at a robotics company but you wouldn't be the best at your job if you didn't understand how the whole system works. Honestly, it's hard to be a good robotics engineer if you don't love robotics and are doing this only for a paycheck. It's possible that this becomes a very big industry and there are people who work only on parts of it, but then you lose the early mover advantage
Yes. There is no compiler to help you out in most tasks, there is no garbage collection to help you out. If the thing fails you have to understand what is happening. Hard transition to make.
@@lemonetrambone I started with programming, then went to neural networks, and now I’m just starting to learn robotics. Initially I could not find information about how people actually got these robots to do free form tasks like picking up squares and sorting them by color, for example. But I would say machine learning is prob the most important part for the whole thing, reinforcement learning in particular.
Thank you for this sober and informed view of this topic. All the hypers, though they may attract many funds to the field right now, should be aware that hypes create expectation bubbles and when they burst, many times a big peroid of questioning and a big drawback can occur.
I almost titled this one "de-hyping AI robotics" lol There's an old book called Crossing the Chasm, I think I do videos on a lot of tech stuff overlaid with that model. Hmm.
Hey Will, just dropped by to say that after watching your video, I'm convinced my toaster is plotting against me with the vacuum cleaner. But seriously, your dive into the robotics and AI world was as enlightening as it was entertaining! It's like we're living in a sci-fi movie, except the robots haven't taken over... yet. Thanks for making such a complex topic accessible and fun. Now, if you'll excuse me, I need to go reassure my household appliances that I still love them, despite my newfound fear of a robot uprising. Keep up the fantastic work, and I can't wait for your next video where you teach us how to negotiate peace treaties with our future robot overlords.
Not sure if you have seen Red Dwarf, but there's a talking toaster gag in I think the first season. Also a famous Rick and Morty joke about a butter fetching robot. :)
Yeah, we are living in the prequel of the Sci-fi movies. What we are experiencing is what's usually covered as a recap in the 1st 3 minutes of the introductory scene of the movie as a recap of how we got to the time period the story is based on. 😂
You’re obviously thoughtful. I liked your video about “when will the developer jobs come back”. Really put it into context why I’d had such a hard time getting my most recent position. Hm, robots, eh? Seems like an easy transition for some SWE’s. I for one already know C++, that’s a start. Well, maybe if all the software gigs disappear, I’ll look into it. I’m still getting paid to push tickets through JIRA for now…
I have been looking for this type of content after taking a break from tech last year. I have a toddler to care for, and the job search is significantly more challenging this time.
IDK, I think a general purpose personal robot would be as transformative as the general purpose personal computer was. You'd download/install behaviors and skills, the hardware is capable of navigating and interacting with the environment. It might not look like a person or be the size of a person, but something the size of a large dog with a hand for a head could get a lot done.
@@bryangrove8275 Like what. I can't think of anything that isn't already automated..washing dishes, sweeping floor, cleaning clothes. I guess the stupid roobt can fold it? After I buy it, fix it, recharge it, find a place to store it, etc? How about i just fk it lol. I don't need it just like I don't need a smartphone.
Both are very reasonable options. If you like the coursework I’d say they are good. Would you still do it if salaries dropped? Do you have a good bead/internship/part time job for summer(s) lined up? My two cents - embrace the ai / ml stuff. Maybe consider a minor in something not computer related that’s a domain you are interested in.
You can just specialize in embedded software and engineering without the need to switch majors. However, switching 100% to CompE is a viable option and less travelled road.
its funny, i predicted this and tried to get ahead of the curve by learning c and computer science [especially logic gates and how they are implementations of boolean functions etc] but i was getting way ahead of myself as my focus currently is web dev. so i'll continue focusing on web dev for now. if i have to go back and learn robotics well, we will have to see. i predict that if ai and robotics really hit the market like corporate america would like, that the world will be so drastically different, that simply going and learning how to work with ai and robotics may not be an easy as an answer that we think. The economy and culture will completely transform and be turned upside down.along with the advent of AI and robotics, new opportunities, new ways of living will emerge that we cannot predict yet. So anyone thinking "i'll just learn robotics or ai" may find they will likely rethink that possibility when AI and robotics truly enter the world.
There were multiple times you said something like " - now, we could spend an entire video discussing that, but..." After the third time you said that, I found my self thinking "You know. I am liking the way this guy talks about stuff. I could look forward to that". Sub'd.
I worked in the manufacturing sector for 10 years as a Design Engineer/Project Manager and at the end of the day, you always have some tyrannical owner/boss looking over your shoulder and breathing down your neck because you have to pull off a miracle or the whole company will collapse. As a result of this, you end up working 15-20 hour days (including weekends) and never leaving the factory. NO THANK YOU.
I would say that robotics hardware is the place to be. That could be electrical or mechanical and aerospace engineering degrees or a combination of. All these degrees give you the fundamentals to pivot into these careers for hardware (and maybe software with a bit of self-learning). I say this because robots require specialized knowledge such as electronics, dynamics and some robots go to space (which is a different set of challenges). So you cannot have a robot without those engineers.
Growing but small? I don't know how many companies are going to be doing this but FWIW I do know that Microsoft is doing this in a big way. Not sure how much automated transformation tools eg guys like this (tsri.com/ - just down the street from me lol) and AI tooling vs manual work.
Hi Will. I've been enjoying your videos. Lots to think about. I'm a self taught web developer and I plan on teaching myself basic robotics using Arduino. Do you think it will be possible to have self taught robotics engineers like we have self taught developers? The barrier to entry seems much higher.
Short answer, sure. As long as you can afford the parts. Check out the open robotics site I referenced in the video Remember that hard to learn = smaller pool of competitors. Imho the hardest part are figuring out which parts to learn and how to translate that to a job. Maybe find some companies in your area and try talking to someone. If you meet with someone at one of those companies, they give you a reading/learning list and you come back in a few months having done it I’ll bet you’re in ftw
While I think we will be moving to smarter manufacturing and more automation, I think the proponents of that aren't facing the consequences directly. The flip answer seems to be, 'retrain manufacturing workers to run automation'. Yeah, some can, but don't you think if they were capable of that they would retain now and make more money already? The rate of change seems to be accelerating, which magnifies problems like worker displacement. I don't have an answer that I think is adequate. There is an upside to all of this -- new, powerful tools are becoming affordable which could allow more entrepreneurs to start small, agile, businesses that were not possible before. Perhaps the new entrepreneurs can hire some of the displaced factory workers if they are willing to re-learn manufacturing.
@@ChangeNode I am in what's probably not too uncommon a situation: I am a trailing-edge boomer with 40+ years experience in IT. I am probably 2 years from retirement, which is likely to be thin enough that I will need a side hustle of some sort. I am willing and able to learn new things, but what kind of side hustle can I do self employed that will leverage my years of UNIX/Linux dev, test, and support experience? I am wanting to move towards embedded systems and robotics, but am wondering how to make this work. Ideas?
@@jeffcauhape6880 off the cuff... Self-employed -> you have to do the marketing/biz dev side yourself. In my area (Seattle) the best way to find contract gigs has been to do things like present at user groups (SeaJUG in my case). I've also been quite surprised at how I've gotten gigs by posting videos on TH-cam on obscure technical topics. View counts kind of don't matter for this - eg I had less than 500 views on a vid that led to two gigs. I'm skeptical of online marketplaces but YMMV. If you don't want to do the biz dev yourself reach out to every IT recruiting shop you can. This book (Nomadic Developer) amzn.to/43eQZIK has been around a while but is still relevant and might be very helpful. Beyond that when it comes to expanding out one option depending on skills/interest might be something like 1:1 tutoring. Around here you can make $50-$75/hour doing stuff like online or in person after school tutoring esp for technical topics. Depending on your target you might be able to increase that with things like 1:1 coaching for college/other tech people.
How do we transition to it without paying for formal training? With software/web, we could self teach. Maybe buy small robotics like Arduino and work your way up?
I think you can go a decent way down that path, but I think the big industrialized stuff use more specialized tooling. So you wind up with a lot of interesting conversations about what/how to pay for that training, software, dev kits, etc. Might try finding middle managers or recruiters in companies of interest and see what they tell you.
Hi Will I'm about 40y now I'm hearing that companies refuse to hire programmers at my age they think we should move to management or CTO or something we shouldn't code they question us, what do you think?
That seems inconsistent with the reality of the work ethic (or lack there of) of the younger crowd. I worried about this myself until I went on temporary assignment working at one of Boeing's campuses in Seattle. Most of the employees were over 40. Many mid to late 50s. In fact young people (< 40) were conspicuously absent.
The age demo data on this one haunts me 2022.stateofjs.com/en-US/demographics/ My take is that there are still a ton of places that are open to hiring a 40yo. But I do think it gets harder. Some verticals are much better eg I found healthcare to be more open. I get the sense that a lot of startups tend to be worst about this. FWIW IMHO the more the company leans on leetcode style interviewing the more likely they are biasing towards recent college grads.
I always wonder how a universal basic income would work considering things like debt, like certain people would have mortgages, student loans etc. Makes me wonder how that would all pan out.
@@motionsick lol if you can't tell from my videos already it's 100% not going to be a panacea vid. Mostly just walking through the $ and what would probably have to happen first (ie at least a great depression level economic collapse) before it would get support. Shh don't tell anyone tho it's a spoiler for the next vid lol
@@ChangeNode oh yeah sure, I mean everyone’s needs and spending would be different, if a UBI came into place and a family had a mortgage, how would that be handled? Would everyone’s debts be reset? Would a UBI be higher for someone with debts?
@@coreydunkin UBI just the government sending out checks to everyone. Think of it like universal social security. I'm planning on breaking down the math for a $500/month check and a $1,200/month, compare to other programs in scale/scope, and talk a bit about how bad the economy (and in what way) to get enough political support to make it happen. Also some about fiscal vs monetary policy.
I’m currently a student community college and I’m in my second semester. Not really knowing what I want to do, I started to take some CS classes which I enjoy. Due to all this AI stuff now though and lay offs, I’m concerned about continuing down the CS path. Should I maybe look into more of a mechanical or mechatronic engineering degree? I would like to do some type of engineering and CS seemed to be hot but after watching some of your videos and seeing the news and stuff, would mechanical engineering be best? I’ve also been looking at some cybersecurity stuff too, how much different would that be from CS?
WRT mechanical vs mechatronic vs others, more or less impossible to predict long term. I'd sort of do a mix of hitting Indeed and LinkedIn hard to see what pops up. Basically really research what's actually hiring. Also of course factor in what you like doing - eg switching from CS to mechanical but oh wait you hate mechanical isn't a good plan. tbh my suspicion is that CS/using computers will be around for quite while but it may very well go through a period of pay reduction. eg that $180k job may drift down to a $120k job or maybe even an $80k job. Who knows? But if that doesn't freak you out then I would say go for it. Just stay flexible and be a tech omnivore.
If you are wanting more flexibility, get a degree in something like computer engineering. It’ll be a much easier transition from CE to CS job than it will be for CS to anything in engineering because you will not have the background in electronics, embedded systems, control systems, and physics to compete or thrive. Just my two cents. Computer Engineering is great because you’ll still learn algorithms, computer architecture, and networking. Embedded doesn’t necessarily have OOP like Python or JS, but we style it in a way it resembles OOP.
Im curious about that statistic of 5-10% of product cost being from labor costs (4:34). Obviously it’s very different depending on the industry and you are talking about the automotive industry and mass produced goods. From my experience in aerospace and avionics electronics, it seems like these industries labor costs are a significantly higher percentage of the final product’s cost. I think the overall “cost savings” across many industries will drop significantly as humanoid robots become more common in the production process. I agree that this will be a huge societal problem with significant job cuts and ever changing economic factors. One interesting side note is tesla’s claim that the cost of a humanoid robot will only be 25k. The actual cost in a few years will probably be closer to 35k because they often overestimate their ability to have such low prices.
The 5-10% figure is kind of the low point. If it's higher it's a bigger target for automation. FWIW I did find some data that seemed to point to a relatively stable labor % component, which made me think that a lot of the automation is improving quality/feature set vs. lowering costs. If the actual cost is, say, closer to 30%+ and the automation can get that down, that's going to be a huge feedback loop. "Buy this car, it's 25% cheaper and made with automation, so the quality is damn near perfect" is one hell of thing to have to compete with...
I agree 100% about scalability. We have great graphene batteries in labs but no way to mass produce them. It turns out that a lot of great inventions never make it to market because they don't scale up well.
What's your take on BlackBerry onx 8.0 is that something that a developer will need to know I hear it's the foundationc micro kernel for nvidia's DriveThor and Jetson combine Blackwell chip stack?
Short answer: no idea. Most of the ROS & robotics platforms feel very small and specialized to me coming from more mainstream stuff like REST/cloud/mobile/desktop. How they are going to combine the various LLM tech w/robotics stacks seems pretty greenfield to me, but the # of job openings seems small.
Interesting. I was planning to transition from edu to tech but this makes it sound like that isn't a great idea now. On the topic of the prefab housing, do you think it would be better than the 3d printing of houses that companies like cybe
WRT switching from edu to tech, honestly most of the folks that I know in edu are absolutely not making enough $ to live comfortably - it's a big problem IMHO but that's another topic. I honestly don't see a lot of fields that aren't at least somewhat concerned about AI/AI robotics, so at some point IMHO it's a matter of just making sure you don't overload w/debt while find something that you both like and has job openings. So do your research on eg indeed to sanity check and network. WRT prefabs, mostly I've been following this channel www.youtube.com/@KerryTarnow and it's been very interesting watching all of these companies trying to sort it out. My guess is that the ideal would be prefab components that are easily shipped. Sort of like big legos but when everything is assembled you couldn't tell. I have seen a few 3d printer options that do things like extruding a concrete like substance, but that's a lot of mass and I'm not sure about the costs & insulation issues. Plus for the 3d printed/extruded there's a lot of manual work running wires/plumbing etc. I think it's all the stuff we don't see because it's usually hidden behind walls that makes it complicated. I guess to put it another way, it's a very dynamic space and very much looking forward to seeing how it shapes up...
Robotics is more mechatronics or mechanical engineering than software engineering, add that programming motions can be taught with motion capture sensors, sure the gaits and trajectories can be recorded and analyzed but just how much actual programming is involved? Robotics is closer to CNC manufacturing than AI text to chat so expect a lot of machine code just to ensure the platform can move about.
The training via example etc stuff is the part that really makes me dubious about the size/scale of the robotics field. IIRC figure and Nvidia are both making a lot of noise about not needed devs for this stuff.
Brilliant, this is exactly what i've been thinking last couple of days, software engineering is on its way to be automated, but AI, and robotics Engineering can still buy some time, all the best insh'Allah :)
I literally spend a little time every day with the K&R C book. I’ve also begun to implement data structures C for practice. So much more fun than JS. Hard (pointers to pointers I’m looking at you), but fun.
Yeah, no kidding. I did a TypeScript project recently and was kind of astonished at how nice it's gotten, but there was something missing about working on it that I just couldn't put my finger one it. TBH one of my personal interests right now is how different stacks handle foreign function interfaces back to C libraries. I honestly don't know why I find it more fun, but I certainly feel cool working on it lol...
@@ChangeNode dealing with memory directly feels like doing something actually tangible. I know it’s fraught with danger and you can bring down the system with just one forgotten pointer, but that’s half the fun
Thanks for your insights. I am 20 years in the industry. Did a lot of Embedded C, wrote network stacks for proprietary OSes running on healthcare devices. I have code in linux kernel. I am also fluent in python and golang. Lastly I have been developing also some custom kubernetes controllers. I am trying to stay AI resilient. Looks like future of system programming is rust and not c/c++ anymore. Would you agree ? Do you have any advice how to learn Rust quickly for a person with this background ? Would it also make sense to add ML skills to this background ? Am I AI safe as a seasoned Embedded C programmer ? I would consider myself as a world class expert in embedded C.
FWIW when I saw Rust in the kernel that really caught my eye www.zdnet.com/article/rust-in-linux-where-we-are-and-where-were-going-next/ - I think MS is also busy writing/rewriting stuff in Rust too. WRT how to learn just depends on you. There are tons of books, TH-cam videos, etc. You are already massively ahead fwiw. I did a quick search and found this github.com/nrc/r4cppp I'd look at Rust first and then look at ML as well but I suspect you'll find it easier to find Rust work...? I think you're likely as AI safe as just about anyone is. For now it's the usual stuff, keep learning, network, talk to recruiters to sanity check etc.
Network network network Check out tools like bubble.io and also check out stuff like lmstudio.ai/ and fold that in to the rest of your skill set Be able to talk to metrics, funnel/pipelines, data processing, statistics, including a very healthy skepticism for cherry picked #s, very common in middle management. But, uh, in a way that doesn't annoy them. Ahem.
Really dig your content. Could you make a video talking about how you see Brain Computer Interfaces (BCIs) advancing with companies pursuing human integration, as well as some robotics companies developing prosthetics for human augmentation? I think I can see these two joining in the next 30 years.
Interesting topic. My hot take is that we are going to see BCI mostly limited to medical cases (places where people have lost function and it helps return) for quite some time. There's something very very squick for most people about implants for some obvious reasons. Whoever does BCI will have to be very, very, very trustworthy. They will have to be all in on partnering with FDA etc. Given all of the changes right now trying to sort out the next ten years seems insanely tricky. 30 years? I mean, try explaining 1960 to someone in 1930 but way more intense...
if they are trained mechanical or electrical engineer, you can pivot. if not you might need to consult engineers. aside from that you are programming low level. not all developers are willing to go back to c and assembly.
Yeah, it's rough. Other folks in robotics are also commenting that it's grim, bad pay, long hours, not a lot of jobs. I don't want to be totally negative - if someone has the dream(tm) of being a robotics engineering I say go for it, but if it's just $ then no...
What do you think of telehandling as way for people in the poorer countries to provide services in richer countries. Could someone in China work as a laborer in US and make financial sense?
the tl;dr version, my guess is that having someone running a waldo (en.wikipedia.org/wiki/Remote_manipulator) isn't that appealing for a variety of reasons or we'd probably see a lot more of it today. Interesting to try to figure out why, although a few reasons do come to mind including fussiness/costs for the robotics. That's part of where industrial automation/robotics is where the $ is I suspect.
@@ChangeNode ye... I would, but I would have to relearn a lot. It's a whole different way of thinking. Different design patterns. Parallel. It's not an easy transition.
The industry isn't structured for the multi disciplinary world of robots. They want specialists, not generalists but building robots requires a lot of generalists. Still waiting for blockchain for robots.
There are way more jobs right now for Java/Spring I think. Longer term you are going to want to get very good at being able to pick up languages fast. Rust and Java are both C/C++ descendents so that part isn't too bad. The tricky thing about a new stack is always the libraries/SDKs. FWIW one of my favorite books amzn.to/4cbUYts even if you don't go through the exercises it's a great read. I'd get a part time job/internship for this summer lined up ASAP. My guess is market will still be soft in 1-2 years but (crossed fingers) will be better after that. Your #1 is to get ~3 years professional experience on your resume.
I actually think the term AI/Robots and similar are used more for marketing hype. Automation is a great tool that makes sense for business, but getting eyes on newspapers and TH-cam videos. For that, you need something more exciting. :)
Oh, depends on the robot. But absolutely different people have very reactions. Same for stuff like motion sickness - eg there are a few apps on my Quest 3 that my son is fine with that make me instantly queasy. My guess is that they a) they will tweak to fit for purpose and/or b) we may just get used to 'em.
Yeah, Rust in the Linux kernel was huge. Under the covers so much stuff still relies on bindings to C functions, so still worth learning at least how to read IMHO.
I think that mostly depends on the role/field, but generically yes. I think you could get a long way doing self-study but the sense I get looking at the job descriptions on Indeed/LinkedIn is that there is just so much to cover I think it would be very hard to break in without a degree. You might want to check out the open source robotics stuff I mentioned in the vid and just get a feel for it yourself.
i was thinking about getting a CS degree but after the latest news i changed my mind and now Iam looking for safest engineering majors from automation and Iam interested in two (mechanical engineering and electrical engineering) so what do you advise me and are those really safer than CS ??
CS degrees are still good. You may hear a lot of doom about how AI is going to replace all software engineers, but most of that is nonsense. The nature of software engineering is definitely going to change as we learn to adapt with the tools, but you will still need software engineers. If AI really does replace software engineers, then every type of engineer is toast lol. If you go to a good school, have a great portfolio, and get good internship experience, you will have nothing to worry about (the current job market is tough, but things will get better within a few years as the market recovers). CS overall is still a very in-demand, fast-growing field leading the frontier of new technologies like robotics, ML/AI, etc. If you don't end up majoring in CS, at the very least learn to program and keep up to date with ML and that will give you an edge in whatever field you end up going into. Good luck!
Electrical engineering is good as AI still needs hardware. Also there’s enough code involved to pivot to developer role fairly easily. If I were doing a degree today I’d take electrical eng.
Sigh. First, if you’ve been using LLMs to write code for you and you’ve concluded that software development is over? It tells me you’re either not a very good developer, or you’ve not used LLMs very much for code. OpenAI, Llama2, Claude. They all make mistakes. A lot of mistakes. They will always make mistakes, because LLMs are NOT REAL AI. This will be a productivity tool for you to use in the future. If companies hired developers to write code for them that you can copy off of StackOverflow. Guess what? They’re going to pay you to copy code from an LLM and make it work. You guys all sound like Boomers in the late 90s when they thought computers were going to take all their jobs. Here we are 25+ years later……
lol yeah. tbh part of the point of the video was to kind of explain just how hard it is. Don't want to discourage anyone but also wanted to explain that it's not actually likely to be easier. Pondering doing a similar one on plumbing, ha
lol started coding when I was 10, BASIC on Commodore PET and Apple //c first paid job was when I was 16, writing macOS stuff in Pascal macOS Pascal + a little bit of 68k assembly in college, working for a lab converting bits of COBOL, FORTRAN, and BASIC to "modern" Pascal. Been going ever since. Apple & Symantec in the 90s. Startup from 1999-2001. Consulting from ~2002-2016... Anyhoo... yeah, paid since 1992, FTE bonafide SV tech since 1995. So, yeah, that's decades. It's completely insane looking back on it...
If we really think about it, with the industry really care if AI robots working in our homes put the cat in the dishwasher? I mean, we have airplanes being flown by human beings falling apart mid air, and nothing is being done about it. You have people who built those planes on camera, saying that they would not fly in those planes.
If Robotics explodes- then so will the proliferation of abstraction/frameworks to interface with them. Very few people will actually be Robotics/Systems Engineers, in fact they will be 2 or 3 levels abstracted away because protection of the source code will be of upmost importance. I think it will create newer DSLs and scripting languages for people to interface with. It will be the same as it ever was. Instead of a funky app on a smart phone ,it will be a funky app on a smart phone to interface with a robot in your house like a roomba.
Definitely disagree with companies being tone deaf to terminology around robotics and ai. None of these companies care about what the public thinks. They want to induce anxiety in investors and they are good at it. If the work a company is doing seems scary and paradigm changing, you better own a piece of it.
He said that you only save about $5000 per car, but that is wrong. Image if robots are used all the way down the line. You have robots harvesting the raw materials, You have robots creating the steel, You have robots building the robots to do all these task. Soon the cars cost is under $1000. With little to no humans any where along the line, that is the future.
Yeah, there's absolutely a line to that eventually. In another thread (on the what is AGI video?) it meandered into a way to measure the capabilities of a robot in terms of how close it can get to full self-construction. Everything from resource extraction, processing, and assembly. Gets murky eg when considering just using existing things like chip fabs vs actually being able to do chip fab. +1 for it being the future, but the timeline is murky...
I am worried that people are looking at this with a cloud of misconception, they think their job is safe until those humanoid robots become common The WFH during the pandemic as I see it now is a double edge sword, to all jobs that could function without being in the office could potentially be replaced by AI in common computers not robots Been looking at this for months, I conclude there will be job reductions in the future and it will be massive, as early as now I am trying to prepare for this financially, physically and emotionally I hope people, in just small steps daily, start preparing for this
The next video on AI Robots cost is more inline with some of your concerns I think. This stuff is moving very fast and it's very hard to keep track. WRT preparing... yeah. It's a very strange time.
@@ChangeNodeI don't know. What I said was over all I agreed with you that tech is more than just software engineering. I shared my experience with the 2000 doot coom bubble and I said I was umemployed for so long I ended up joining the Army out of desperation. I just thought you (or someone else) thought I was being pro-military and trigger the comment delete...
@@bigbigdog no idea, wasn't me. FWIW I did check and there were two comments held by TH-cam auto mod for review, but they weren't yours and I just approved both of them. AFAIK you can try resubmitting and see what happens. Did TH-cam send you an email or anything? If not my guess it that it just got eaten by the server or some connectivity thing...?
@@ChangeNodehm... Thanks for checking. TH-cam has been acting weird for me lately. It's ok. The other comment was pretty long and I kinda forgot what I worte anyways. tks.
I don't think it's a good idea. First, only very niche hardware spaces are still okay: FPGA, ASIC... but forget about GPU, CPU, and above all MCUs. These are massively oversupplied fields with thousands of applicants only from India, Pakistan, Indonesia, Philippines etc. Second, most software developers need years to become proficient on hw.
Yeah, most of the robotics/hw folks responding in the comments have all basically said "don't" Part of what I wanted to do was gently explain to folks how much work it is. Kind of like being a doctor or electrician, I think a lot of folks don't realize how hard/difficult it is, but if you love it you love it...
I would like to see japanese style vending machines and robotic restaurants. for 2 big reasons. I don't about where you live, by my experience with fast food is: it's overpriced and the people who make it are NASTY.!!!!! because of that, i no longer eat at restaurants.
Funny how you mention the investments going into 1X/Figure but leave Tesla's Optimus out lol.. Mass manufacturing is what matters and Tesla will own on that front
My dentist had what was technically a CNC system that would cut a tooth or cap 15-30 minutes based on a scan. He'd then order out for the full version but it could take days. I don't think it was that expensive and it sure made life a lot better as an interim solution. FWIW.
Your math on replacing labor might save 10% doesn’t account for the humans replaced in the supply chain, so it’s flawed math. You need to look at all the materials used to build cars or the items on shelves in retail. It’s not just the retail staff that matters. It’s the savings in the entire supply chain.
Oh, absolutely agree. There's a bootstrap test for robotics - how much of the entire supply chain can you automate with a robot? Part of what I am focusing on right now is the transition period and minimum economic viability of this stuff. Once it starts getting any deployment I think it starts becoming a self-reinforcing loop...
Nah, better build out cognitive assistants. There is no shortage of physical workers, but cognitive workers are very rare. We got a lot of actors impersonating cognitive workers, but that’s quickly going away.
I'm sure you've seen this demo by now? th-cam.com/video/Sq1QZB5baNw/w-d-xo.html How might a developer pivot into robotics without ready access to great hardware?
RE figure, between that and Devin dropping essentially right after my vid looks like I've got another vid to do lol RE: robotics a lot of the controllers/subcontrollers look like the are driven by tiny little Linux hardware, you can get pretty far with C/C++/Rust on a comparative potato. Local LLM stuff is a lot more about the big box, although I have played with a few sub 2gb LLMs that could run on my iPhone. Check out the open source robotics page and you could find some pretty inexpensive stuff to play around with.
@@ChangeNode Thanks for the helpful response. Really appreciate it. And yes, by all means, please post a follow up video. Would love to hear what you have to say. Subscribed and staying tuned...
Energy is the answer, with cheap energy more things are possible including make water/food from the air. AI/automation needs lots of energy, solar/wind/batteries are not the solution.
I don't even know anymore www.reuters.com/technology/microsoft-buy-power-nuclear-fusion-company-helion-2023-05-10/ www.theverge.com/2023/9/26/23889956/microsoft-next-generation-nuclear-energy-smr-job-hiring It's getting awfully cyberpunk out there
Robotics sucks. I did my masters in robotics, and tbh the pay in that field is 1/4 what it is other places. Meaning, specifically, similar roles pay literally 1/4th as much. I could not justify staying in the industry so I changed to Web 3 and edge computing (solving similar problems) and I now make 4x the salary.
@@ChangeNode EVM work. Its quite easy for me to find work at least, and I do get messages from recruiters 2-3 times a month. Cant predict the future tho!
Would you agree that it's a good thing that Earths population is on the decline, just as we enter the age of AGI powered androids and a decreasing need for workers?
RE: Earth population & demographics, that's a big topic and it starts to get into a lot of cultural, political and economic complexity. So... I'm going to punt on this one mostly for now. I guess I will say that if a country / system can't manage to maintain population levels, eventually it will disappear, which seems less than ideal. Most of the objections to growing population seem to be rooted in climate and resource impact. I personally do think that there's no reason we can't dramatically reduce the resource impact for individuals assuming there is system wide support. If we can't figure that one out... well, I guess that in the long term that's yet another answer to the Fermi paradox. Sigh. FWIW en.wikipedia.org/wiki/Malthusianism
Your comment that a Restaurant becomes a glorified vending machine, so why would anyone want to go is missing the point of restaurants. We don’t go for the servers. We go, so we can have a nice prepared meal we don’t have to cook and we don’t have to clean up after. It’s not for the human servers.
Yeah, you can break down the components of going out in several ways. Human contact? Ambience? Don't have to cook? My point was more that for a certain class of restaurant, just having the experience turn into a vending machine was missing the point, whichever of those components is most important. I have seen videos of sushi restaurants in Japan that are basically entirely served via order screens + a conveyor belt. I think someone still comes in to clean those up by hand. Either way, my guess is that the human contact is the last thing to go. If you had a figure.ai robot that would prep/cook/clean up then why go out? My take is human contact.
@@ChangeNode I guess it depends on the type of restaurant. Fast food or Starbucks etc is primarily cost, convenience and taste. I don’t care if I’m served by a robot if the other 3 are good enough If it’s a Chop House looking for a great dinner it’s about food and ambiance. Convenience doesn’t matter much. Cost is what it is…. A great server is a nice to have, but often their value add is getting our orders right and quickly. If a robot “could” do both of those things and at a lower cost I wouldn’t complain. 😊
What should my next video be about?
Is blockchain officially dead?
Developers: Pivot to farming? 😅
Recently laid off Amazon Robotics Engineer, I think your skepticism is spot on, some of the skillsets of a developer can translate well. The problem is these big companies often paint a grandiose picture of their robotics for press release and rarely deliver, usually do to economic or Hardware issues. Hardware trails software, and you have to take a hardware first approach as Steve jobs mentioned.
In addition, companies like amazon will have 2-3 separate projects that essentially do the same thing and compete internally so they can hedge their bets. The engineers on the "losing" teams often end up looking for work elsewhere, just look at projects like Scout.
How are you feeling about another robotics gig? Do you think it's easier or harder than, say, a Java REST dev?
Correct. Right now the center of hw & electronics in the US is SXSW & Mexico. But it still trails significantly to South East Asia. But the Americas is starting to bring that back. It will be exciting to see what it looks like in 10 years.
@@JRay2113 Good to know. Thanks for that.
Firing engineers for non-engoneering mistakes is stupid and wasteful.
2-3 teams competing on different systems design is fine but that's a special role who makes those architectural decisions.
Software/mechanical engineers execute those designs and must be fired only if they fail in their own engineering crafts.
Yup, and you gotta know your hardware and circuits at least at a basic level
I've been watching a lot of low-level programming channels lately and started dabbling with C. Systems engineering is a field that I would like to get into soon... sounds way cooler than building the 100th version of the same crud app.
Yeah, no kidding. And the crud app is exactly the bread and butter for the low/no code world.
Honestly I really don't like low/no code as a dev all that much. But I finally sat down and learned bubble and 1) it certainly can build a LOT of stuff 2) a lot of what it can't do can be added via just a few custom REST services or whatever. Can't do everything but it sure can do a lot.
Not sure where you live but if you are a decent handyman in Seattle you can make way way better than minimum wage. FWIW.
Im a decent handyman in Seattle and you are correct.
😂@@motionsick
@@yehororlov8362 that'd probably be a good bet considering that handyman jobs are going to be the best jobs in a couple of years time 🙃
I'm a robotics and AI engineer, with a background in Mechanical engineering and went to grad school for robotics and AI. It's not an easy pivot or a career to go and chill 😅. Robotics of previous decades is not the same as modern robotics. Before, you needed to understand electromechanical fundamentals and you could work your way through the rest. In today's robotics, you need to understand some new fundamentals you need to have development skills to start with then add on the fundamentals of AI and background knowledge in computer science. Then you need to choose your focus whether it's manipulators, Mobility locomotion, Sensing and perception, Operator interface, Manipulators, and effectors each of those requires programming skills. The same computer science fundamentals are required when building autonomous systems (could be vehicles or some other system). It's also not that easy for the companies building these systems to make them work and produce them in mass at a profitable cost and retain their staff. It's a long-term investment career and I would consider it as R&D. You can also go into automation & manufacturing which is more accessible and stable as a career since manufacturing plants use automation anyway.
Thanks for the details!
Would you be open/interested in doing an online interview about what it takes to get into the field and maybe your thoughts on the big $ investments in the news? If so drop me your contact info changenode.com/contact/ :)
thank you for your comment! Im a student doing my Master in robotics and autonomous systems. it was pretty tough for me and i still struggle to image where I can work in the future. How to find out?
@@xptransformation3564 your first step IMHO is to start calling/networking with recruiters in your area and see what they have to say. Your education is an excellent first step - you need to fold in talking with hiring managers ASAP. Good luck!
20 years in IT here, expertise in development, masters in computer science & AI degree. On each AI job there are 100 applicants within an hour. Recruiter told me 400+ Media make rumors that does not seem to have translation in reality. btw I was laid off because it was no AI work any more. That's why I doubt all these 'exciting' news are reliable.
@@rafae5902 Depends on the job, eg I know someone who works on Android Automotive and someone else who does IoT stuff all from home, but for a bunch of it I would think you'd have to go in. The devices for both of those guys are pretty small, so no biggie to set up a dev desk at home. I could see where for some of the other stuff it might be work hybrid, but still have to go in at least 1-2 days a week.
as an agriculture engineer and web developper, i think Agriculture is one of the sectors where robots will make a huge impact
I like your format, no hype just a conversation. You've clearly thought about this and you have context to see the bigger picture.
Exactly
Thanks! :)
Robotics engineer here - labour cost alone is not a meaningful metric. Human presence reduces opportunities to automate (a lot of safety-related limitations). Even if cost is only 10% now, removing a human altogether can allow you to build cars a few times faster, so cost savings will be way above 10% that you'll shave off human labour.
And don't hold your breath on robots co-existing with humans. Even if technology will deliver (spoiler alert - it won't), the regulations will catch up years later. Large red e-stop buttons will be pervasive for years to come.
As for 3D-printing - not that useless, 3D-printing unlocked generative design. You cannot machine such shapes with the regular subtractive machining, but metal 3D-printing can do all those organic shapes easily. In prototypes or in scale - does not matter.
Underrated comment with someone who has actual industry experience
Yeah, tl;dr I agree with all of this over the long term. The Volvo plant is basically new, so that's why I used it at an example of what exists today. Any thought on timeline for that kind of factory going dark?
RE 3d printing in particular there's a bit of a terminology issue. I'm kind of short handing 3d printing as the home units. Technically I'm sort of confusing/labeling an industrial scale 3d printer as industrial automation. The point I was going for is that we don't really do 3d printing at scale at home.
Thanks!
Thanks for that contribution. Fascinating bit about generative design.
Maybe the real problem to solve is how do you introduce robots with humans present and still get the benefits without sacrificing safety. The Army is funding research for drone medevacs, where an unmanned helicopter carries a patient from a remote location. The solution isn't to get rid of humans so that robots can do their jobs more efficiently.
@@cryora and this is an incredibly hard problem to solve. Soft robotics (e.g., pneumatic actuators) can be promising, but there is a lot of unsolved issues with it. The usual hydraulic or direct drive electric actuation is not yet suitable for operating in the same space as humans - even if sensing and control is solved, there is still a need to prove that it is 100% perfect and glitches won't happen. Regulators at the moment only accept the full shutdown, no actuators powered as a safety threshold for humans (e.g., in industrial robotic cells).
The chasm between demos and production deployments for a robotics system is immense. Far greater than anyone would assume.
I've seen software demos that were literally nothing but made up screenshots/animations purported to be shipping software.
What's your take on the figure.ai demo & timeline...?
@@ChangeNode Very hard to say, but I would say their biggest hurdle will be regulation as their bot is technically a cobot. I'm my opinion configurable Microfactories are a more realistic medium-term solution.
As someone who wanted to get into robotics from the start but no one considered me because I didn't have an engineering degree, I see this as an absolute win.
But I am not gonna apply to the Amazons and what not. I am thinking more the small start-ups.
I didn't knew about this channel. Instant favorite. Thank you for this awesome content (I'm a Bsc in EE, transitioning to Machine Learning and Robotics through masters and freelancing).
Truer words have never been spoken 🎉
I have been looking at Zig as a replacement for C/C++ and has a much lower learning curve compared to Rust, at least for me. Looking forward to how the robotics industry shapes up
Interesting. Just took at look at blog.logrocket.com/comparing-rust-vs-zig-performance-safety-more/ and the comments about FFI caught my eye. I've heard more about Rust in the context of the Linux kernel, and I've played with it a bit in the context of Tauri. Haven't seen Zig come up.
Zig is cool, but I have been using Free Pascal and love the way it forces stronger code typing, Ada would be even better. Not sure why Ada is not used more.
If you are interested in robotics. Learn Rust, it's more widely used in robotics. At least in the Context of ROS.
This is the first video of your channel that I've seen after watching enough other types of tech commentary that 90% of the time just revolves around which language is better, which editor is better, this tired old measurement of tools instead of products, achievements and good investment of time, I've almost slam shut my tablet after hearing how 80 hours per week is the way to go while 50 is enough to maintain sanity. So this mature analysis of the state of affairs, and proper weighing more real-world aspects of the economics behind re-training the workforce, the natural resource dictatorships bent on violent expansion occasionally, which I've mentioned in discussions myself is so refreshing. Thank you for this video!
On manufacturing automation: 4:55
I think the 5-10% labour cost is likely for current huge operations benefitting from mass production/economies of scale. Labour cost as a percentage is far higher on smaller, more bespoke production, so that’s where I see the more impact coming eventually, as automation costs drop.
Cheaper, more capable automation is likely to mean more flexible ‘ cheaper operation on smaller scales, allowing more customised/variable production, faster turnaround & lead times, and shorter production runs becoming more feasible. That could open the door for different kinds of [re]manufacturing on a local scale.
Thanks, this video made me think. I've been an enterprise developer for almost 15 years using mostly scripting language. Last year I took a 10-month C++ course. It was quite an awesome experience to manage memory yourself. Hoping to find a way to switch my career to something more challenging like embedded, or just working on more profound software products for desktop/servers, like OSs, browsers, game engines. Robotics field looks intriguing, but I'm afraid I lack electronics/math skills to get such kind of jobs.
Love your take on the idea. If I want human contact, I want human contact, if I want a coffee from a robot, I'll use a vending machine. :) Simple.
Robotics is a multi-disciplinary field, you need knowledge of embedded systems, low level programming, AI/ML, control theory, path planning and estimation, mechatronics and a lot of maths. You can specialize in one of these domains and still get work at a robotics company but you wouldn't be the best at your job if you didn't understand how the whole system works.
Honestly, it's hard to be a good robotics engineer if you don't love robotics and are doing this only for a paycheck. It's possible that this becomes a very big industry and there are people who work only on parts of it, but then you lose the early mover advantage
Yes. There is no compiler to help you out in most tasks, there is no garbage collection to help you out. If the thing fails you have to understand what is happening. Hard transition to make.
I'd say it is harder to live paycheck to paycheck with a minimum wage job for 40 years
Knew i wasnt dumb trying to figure out to to get into robotics. This feels like early internet era for robotics. No HTTPS standard for the internet
Please share your journey, what was your learning path? 🙏
@@lemonetrambone I started with programming, then went to neural networks, and now I’m just starting to learn robotics. Initially I could not find information about how people actually got these robots to do free form tasks like picking up squares and sorting them by color, for example. But I would say machine learning is prob the most important part for the whole thing, reinforcement learning in particular.
Thank you for this sober and informed view of this topic. All the hypers, though they may attract many funds to the field right now, should be aware that hypes create expectation bubbles and when they burst, many times a big peroid of questioning and a big drawback can occur.
I almost titled this one "de-hyping AI robotics" lol
There's an old book called Crossing the Chasm, I think I do videos on a lot of tech stuff overlaid with that model. Hmm.
Really enjoyed listening to you! Nice tone, format, and content!
Thanks a lot for your vulnerable, honest, non-click-bait videos about these important topics!
Hey Will, just dropped by to say that after watching your video, I'm convinced my toaster is plotting against me with the vacuum cleaner. But seriously, your dive into the robotics and AI world was as enlightening as it was entertaining! It's like we're living in a sci-fi movie, except the robots haven't taken over... yet. Thanks for making such a complex topic accessible and fun. Now, if you'll excuse me, I need to go reassure my household appliances that I still love them, despite my newfound fear of a robot uprising. Keep up the fantastic work, and I can't wait for your next video where you teach us how to negotiate peace treaties with our future robot overlords.
Not sure if you have seen Red Dwarf, but there's a talking toaster gag in I think the first season. Also a famous Rick and Morty joke about a butter fetching robot.
:)
I’m more worried about my RTX GPUs in my house taking me out before my smart appliances get a chance to do it 😂
@@FirstNullLast just remember that plastic burning smell just means that the overclock is working 😆
Yeah, we are living in the prequel of the Sci-fi movies. What we are experiencing is what's usually covered as a recap in the 1st 3 minutes of the introductory scene of the movie as a recap of how we got to the time period the story is based on. 😂
this sounds like it was written by ChatGPT
You’re obviously thoughtful. I liked your video about “when will the developer jobs come back”. Really put it into context why I’d had such a hard time getting my most recent position. Hm, robots, eh? Seems like an easy transition for some SWE’s. I for one already know C++, that’s a start. Well, maybe if all the software gigs disappear, I’ll look into it. I’m still getting paid to push tickets through JIRA for now…
lol move those JIRA tickets! Just make sure to sock some $ away :)
im a better developer than you so you probably wont get that position
I have been looking for this type of content after taking a break from tech last year. I have a toddler to care for, and the job search is significantly more challenging this time.
at least you have a husband who can take care of you. we men must work regardless.
A humanoid robot in the home sounds like nothing but trouble. It seems like smaller robots to do specific tasks would make more sense.
I'd like 100 humanoid robots 6 inches tall to do all my work like little gnomes....helps knowing I can take them in a fight if it comes to that.
IDK, I think a general purpose personal robot would be as transformative as the general purpose personal computer was. You'd download/install behaviors and skills, the hardware is capable of navigating and interacting with the environment. It might not look like a person or be the size of a person, but something the size of a large dog with a hand for a head could get a lot done.
@@bryangrove8275 Like what. I can't think of anything that isn't already automated..washing dishes, sweeping floor, cleaning clothes. I guess the stupid roobt can fold it? After I buy it, fix it, recharge it, find a place to store it, etc? How about i just fk it lol. I don't need it just like I don't need a smartphone.
Another option is Computer Networking, which can be quite complex at the CCIE level.
Thank you so much, your knowledge and insights are very inspiring!! love your content
Very informative; thank you.
Sir, Your adherence to the WEF talking points is commendable!
lol I wish they would send checks or something
Have a different take?
As a third year computer science major, do you think it’d be wise to switch into something more flexible such as computer engineering?
Both are very reasonable options. If you like the coursework I’d say they are good. Would you still do it if salaries dropped? Do you have a good bead/internship/part time job for summer(s) lined up?
My two cents - embrace the ai / ml stuff. Maybe consider a minor in something not computer related that’s a domain you are interested in.
You can just specialize in embedded software and engineering without the need to switch majors. However, switching 100% to CompE is a viable option and less travelled road.
its funny, i predicted this and tried to get ahead of the curve by learning c and computer science [especially logic gates and how they are implementations of boolean functions etc] but i was getting way ahead of myself as my focus currently is web dev. so i'll continue focusing on web dev for now. if i have to go back and learn robotics well, we will have to see. i predict that if ai and robotics really hit the market like corporate america would like, that the world will be so drastically different, that simply going and learning how to work with ai and robotics may not be an easy as an answer that we think. The economy and culture will completely transform and be turned upside down.along with the advent of AI and robotics, new opportunities, new ways of living will emerge that we cannot predict yet. So anyone thinking "i'll just learn robotics or ai" may find they will likely rethink that possibility when AI and robotics truly enter the world.
yeah i saw this a bit 4Years ago coming so i thougt it was a good investment
There were multiple times you said something like " - now, we could spend an entire video discussing that, but..." After the third time you said that, I found my self thinking "You know. I am liking the way this guy talks about stuff. I could look forward to that". Sub'd.
Yeah, I'm trying to keep these to less than 20 minutes, ideally I think 10-15.
Any particular one that jumped out as a good one to cover...?
I worked in the manufacturing sector for 10 years as a Design Engineer/Project Manager and at the end of the day, you always have some tyrannical owner/boss looking over your shoulder and breathing down your neck because you have to pull off a miracle or the whole company will collapse.
As a result of this, you end up working 15-20 hour days (including weekends) and never leaving the factory.
NO THANK YOU.
Excellent content. subbed.
Thanks!
You should directly link sites that you show in the description 😊
You are correct! I’ll add it to my checklist. 👍
People: "Do you even need to know any math for programming?!?!"
Robotics Programmers: "Yes."
I have an extremely critical, very important question about this video. Where can I find more videos of that adorable kitten?
I would say that robotics hardware is the place to be. That could be electrical or mechanical and aerospace engineering degrees or a combination of. All these degrees give you the fundamentals to pivot into these careers for hardware (and maybe software with a bit of self-learning). I say this because robots require specialized knowledge such as electronics, dynamics and some robots go to space (which is a different set of challenges). So you cannot have a robot without those engineers.
Hi I was thinking about the niche of translating c/c++ to rust. Do you think there would be a growing market for this?
Growing but small? I don't know how many companies are going to be doing this but FWIW I do know that Microsoft is doing this in a big way. Not sure how much automated transformation tools eg guys like this (tsri.com/ - just down the street from me lol) and AI tooling vs manual work.
Hi Will.
I've been enjoying your videos. Lots to think about.
I'm a self taught web developer and I plan on teaching myself basic robotics using Arduino.
Do you think it will be possible to have self taught robotics engineers like we have self taught developers? The barrier to entry seems much higher.
Short answer, sure. As long as you can afford the parts. Check out the open robotics site I referenced in the video
Remember that hard to learn = smaller pool of competitors.
Imho the hardest part are figuring out which parts to learn and how to translate that to a job. Maybe find some companies in your area and try talking to someone. If you meet with someone at one of those companies, they give you a reading/learning list and you come back in a few months having done it I’ll bet you’re in ftw
@@ChangeNode appreciate the reply. Thanks
While I think we will be moving to smarter manufacturing and more automation, I think the proponents of that aren't facing the consequences directly. The flip answer seems to be, 'retrain manufacturing workers to run automation'. Yeah, some can, but don't you think if they were capable of that they would retain now and make more money already? The rate of change seems to be accelerating, which magnifies problems like worker displacement. I don't have an answer that I think is adequate.
There is an upside to all of this -- new, powerful tools are becoming affordable which could allow more entrepreneurs to start small, agile, businesses that were not possible before. Perhaps the new entrepreneurs can hire some of the displaced factory workers if they are willing to re-learn manufacturing.
Yeah, the displacement is going to be rough. I think that's part of why so many tech people are talking about UBI...
@@ChangeNode I am in what's probably not too uncommon a situation: I am a trailing-edge boomer with 40+ years experience in IT. I am probably 2 years from retirement, which is likely to be thin enough that I will need a side hustle of some sort. I am willing and able to learn new things, but what kind of side hustle can I do self employed that will leverage my years of UNIX/Linux dev, test, and support experience? I am wanting to move towards embedded systems and robotics, but am wondering how to make this work.
Ideas?
@@jeffcauhape6880 off the cuff...
Self-employed -> you have to do the marketing/biz dev side yourself. In my area (Seattle) the best way to find contract gigs has been to do things like present at user groups (SeaJUG in my case). I've also been quite surprised at how I've gotten gigs by posting videos on TH-cam on obscure technical topics. View counts kind of don't matter for this - eg I had less than 500 views on a vid that led to two gigs. I'm skeptical of online marketplaces but YMMV.
If you don't want to do the biz dev yourself reach out to every IT recruiting shop you can. This book (Nomadic Developer) amzn.to/43eQZIK has been around a while but is still relevant and might be very helpful.
Beyond that when it comes to expanding out one option depending on skills/interest might be something like 1:1 tutoring. Around here you can make $50-$75/hour doing stuff like online or in person after school tutoring esp for technical topics. Depending on your target you might be able to increase that with things like 1:1 coaching for college/other tech people.
@@ChangeNode Thank you very much!
Trunks is already spinning up his time machine
How do we transition to it without paying for formal training? With software/web, we could self teach. Maybe buy small robotics like Arduino and work your way up?
I think you can go a decent way down that path, but I think the big industrialized stuff use more specialized tooling. So you wind up with a lot of interesting conversations about what/how to pay for that training, software, dev kits, etc. Might try finding middle managers or recruiters in companies of interest and see what they tell you.
Hi Will I'm about 40y now I'm hearing that companies refuse to hire programmers at my age they think we should move to management or CTO or something we shouldn't code they question us, what do you think?
That seems inconsistent with the reality of the work ethic (or lack there of) of the younger crowd.
I worried about this myself until I went on temporary assignment working at one of Boeing's campuses in Seattle. Most of the employees were over 40. Many mid to late 50s. In fact young people (< 40) were conspicuously absent.
The age demo data on this one haunts me
2022.stateofjs.com/en-US/demographics/
My take is that there are still a ton of places that are open to hiring a 40yo. But I do think it gets harder. Some verticals are much better eg I found healthcare to be more open. I get the sense that a lot of startups tend to be worst about this.
FWIW IMHO the more the company leans on leetcode style interviewing the more likely they are biasing towards recent college grads.
@@rtwas thank you, It's refreshing to read that
I always wonder how a universal basic income would work considering things like debt, like certain people would have mortgages, student loans etc. Makes me wonder how that would all pan out.
Great Q - can you elaborate more on what you mean?
Stop dreaming about UBI. 😂
@@motionsick lol if you can't tell from my videos already it's 100% not going to be a panacea vid. Mostly just walking through the $ and what would probably have to happen first (ie at least a great depression level economic collapse) before it would get support.
Shh don't tell anyone tho it's a spoiler for the next vid lol
@@ChangeNode oh yeah sure, I mean everyone’s needs and spending would be different, if a UBI came into place and a family had a mortgage, how would that be handled? Would everyone’s debts be reset? Would a UBI be higher for someone with debts?
@@coreydunkin UBI just the government sending out checks to everyone. Think of it like universal social security. I'm planning on breaking down the math for a $500/month check and a $1,200/month, compare to other programs in scale/scope, and talk a bit about how bad the economy (and in what way) to get enough political support to make it happen. Also some about fiscal vs monetary policy.
I’m currently a student community college and I’m in my second semester. Not really knowing what I want to do, I started to take some CS classes which I enjoy. Due to all this AI stuff now though and lay offs, I’m concerned about continuing down the CS path. Should I maybe look into more of a mechanical or mechatronic engineering degree? I would like to do some type of engineering and CS seemed to be hot but after watching some of your videos and seeing the news and stuff, would mechanical engineering be best? I’ve also been looking at some cybersecurity stuff too, how much different would that be from CS?
WRT mechanical vs mechatronic vs others, more or less impossible to predict long term. I'd sort of do a mix of hitting Indeed and LinkedIn hard to see what pops up. Basically really research what's actually hiring. Also of course factor in what you like doing - eg switching from CS to mechanical but oh wait you hate mechanical isn't a good plan.
tbh my suspicion is that CS/using computers will be around for quite while but it may very well go through a period of pay reduction. eg that $180k job may drift down to a $120k job or maybe even an $80k job. Who knows? But if that doesn't freak you out then I would say go for it. Just stay flexible and be a tech omnivore.
If you are wanting more flexibility, get a degree in something like computer engineering. It’ll be a much easier transition from CE to CS job than it will be for CS to anything in engineering because you will not have the background in electronics, embedded systems, control systems, and physics to compete or thrive. Just my two cents. Computer Engineering is great because you’ll still learn algorithms, computer architecture, and networking.
Embedded doesn’t necessarily have OOP like Python or JS, but we style it in a way it resembles OOP.
Im curious about that statistic of 5-10% of product cost being from labor costs (4:34). Obviously it’s very different depending on the industry and you are talking about the automotive industry and mass produced goods. From my experience in aerospace and avionics electronics, it seems like these industries labor costs are a significantly higher percentage of the final product’s cost. I think the overall “cost savings” across many industries will drop significantly as humanoid robots become more common in the production process. I agree that this will be a huge societal problem with significant job cuts and ever changing economic factors.
One interesting side note is tesla’s claim that the cost of a humanoid robot will only be 25k. The actual cost in a few years will probably be closer to 35k because they often overestimate their ability to have such low prices.
The 5-10% figure is kind of the low point. If it's higher it's a bigger target for automation. FWIW I did find some data that seemed to point to a relatively stable labor % component, which made me think that a lot of the automation is improving quality/feature set vs. lowering costs.
If the actual cost is, say, closer to 30%+ and the automation can get that down, that's going to be a huge feedback loop. "Buy this car, it's 25% cheaper and made with automation, so the quality is damn near perfect" is one hell of thing to have to compete with...
Great video. Keep up the good work.
Thanks! :)
I agree 100% about scalability. We have great graphene batteries in labs but no way to mass produce them. It turns out that a lot of great inventions never make it to market because they don't scale up well.
What's your take on BlackBerry onx 8.0 is that something that a developer will need to know I hear it's the foundationc micro kernel for nvidia's DriveThor and Jetson combine Blackwell chip stack?
Short answer: no idea.
Most of the ROS & robotics platforms feel very small and specialized to me coming from more mainstream stuff like REST/cloud/mobile/desktop. How they are going to combine the various LLM tech w/robotics stacks seems pretty greenfield to me, but the # of job openings seems small.
Interesting. I was planning to transition from edu to tech but this makes it sound like that isn't a great idea now. On the topic of the prefab housing, do you think it would be better than the 3d printing of houses that companies like cybe
WRT switching from edu to tech, honestly most of the folks that I know in edu are absolutely not making enough $ to live comfortably - it's a big problem IMHO but that's another topic. I honestly don't see a lot of fields that aren't at least somewhat concerned about AI/AI robotics, so at some point IMHO it's a matter of just making sure you don't overload w/debt while find something that you both like and has job openings. So do your research on eg indeed to sanity check and network.
WRT prefabs, mostly I've been following this channel www.youtube.com/@KerryTarnow and it's been very interesting watching all of these companies trying to sort it out.
My guess is that the ideal would be prefab components that are easily shipped. Sort of like big legos but when everything is assembled you couldn't tell. I have seen a few 3d printer options that do things like extruding a concrete like substance, but that's a lot of mass and I'm not sure about the costs & insulation issues. Plus for the 3d printed/extruded there's a lot of manual work running wires/plumbing etc. I think it's all the stuff we don't see because it's usually hidden behind walls that makes it complicated.
I guess to put it another way, it's a very dynamic space and very much looking forward to seeing how it shapes up...
How do you land a robotics job?
Robotics is more mechatronics or mechanical engineering than software engineering, add that programming motions can be taught with motion capture sensors, sure the gaits and trajectories can be recorded and analyzed but just how much actual programming is involved? Robotics is closer to CNC manufacturing than AI text to chat so expect a lot of machine code just to ensure the platform can move about.
The training via example etc stuff is the part that really makes me dubious about the size/scale of the robotics field. IIRC figure and Nvidia are both making a lot of noise about not needed devs for this stuff.
Brilliant, this is exactly what i've been thinking last couple of days, software engineering is on its way to be automated, but AI, and robotics Engineering can still buy some time, all the best insh'Allah :)
I literally spend a little time every day with the K&R C book. I’ve also begun to implement data structures C for practice. So much more fun than JS. Hard (pointers to pointers I’m looking at you), but fun.
Yeah, no kidding. I did a TypeScript project recently and was kind of astonished at how nice it's gotten, but there was something missing about working on it that I just couldn't put my finger one it.
TBH one of my personal interests right now is how different stacks handle foreign function interfaces back to C libraries. I honestly don't know why I find it more fun, but I certainly feel cool working on it lol...
@@ChangeNode dealing with memory directly feels like doing something actually tangible. I know it’s fraught with danger and you can bring down the system with just one forgotten pointer, but that’s half the fun
@@ChangeNode it’s like we’re junkies; years ago syntax and logic gave us a rush, but now we need the HARD shit
@@rdubb77 😂
Thanks for your insights. I am 20 years in the industry. Did a lot of Embedded C, wrote network stacks for proprietary OSes running on healthcare devices. I have code in linux kernel. I am also fluent in python and golang. Lastly I have been developing also some custom kubernetes controllers. I am trying to stay AI resilient.
Looks like future of system programming is rust and not c/c++ anymore. Would you agree ? Do you have any advice how to learn Rust quickly for a person with this background ? Would it also make sense to add ML skills to this background ? Am I AI safe as a seasoned Embedded C programmer ? I would consider myself as a world class expert in embedded C.
FWIW when I saw Rust in the kernel that really caught my eye www.zdnet.com/article/rust-in-linux-where-we-are-and-where-were-going-next/ - I think MS is also busy writing/rewriting stuff in Rust too.
WRT how to learn just depends on you. There are tons of books, TH-cam videos, etc. You are already massively ahead fwiw. I did a quick search and found this github.com/nrc/r4cppp
I'd look at Rust first and then look at ML as well but I suspect you'll find it easier to find Rust work...?
I think you're likely as AI safe as just about anyone is. For now it's the usual stuff, keep learning, network, talk to recruiters to sanity check etc.
Just about to graduate with a bi major in finance and marketing and a passion for tech, what do you recommend to a person like me?
Network network network
Check out tools like bubble.io and also check out stuff like lmstudio.ai/ and fold that in to the rest of your skill set
Be able to talk to metrics, funnel/pipelines, data processing, statistics, including a very healthy skepticism for cherry picked #s, very common in middle management. But, uh, in a way that doesn't annoy them. Ahem.
Really dig your content. Could you make a video talking about how you see Brain Computer Interfaces (BCIs) advancing with companies pursuing human integration, as well as some robotics companies developing prosthetics for human augmentation? I think I can see these two joining in the next 30 years.
Interesting topic. My hot take is that we are going to see BCI mostly limited to medical cases (places where people have lost function and it helps return) for quite some time.
There's something very very squick for most people about implants for some obvious reasons. Whoever does BCI will have to be very, very, very trustworthy. They will have to be all in on partnering with FDA etc.
Given all of the changes right now trying to sort out the next ten years seems insanely tricky. 30 years? I mean, try explaining 1960 to someone in 1930 but way more intense...
Quite insightful
if they are trained mechanical or electrical engineer, you can pivot. if not you might need to consult engineers. aside from that you are programming low level. not all developers are willing to go back to c and assembly.
Yeah, it's rough. Other folks in robotics are also commenting that it's grim, bad pay, long hours, not a lot of jobs.
I don't want to be totally negative - if someone has the dream(tm) of being a robotics engineering I say go for it, but if it's just $ then no...
TinyML is pretty fun too.
What do you think of telehandling as way for people in the poorer countries to provide services in richer countries. Could someone in China work as a laborer in US and make financial sense?
the tl;dr version, my guess is that having someone running a waldo (en.wikipedia.org/wiki/Remote_manipulator) isn't that appealing for a variety of reasons or we'd probably see a lot more of it today. Interesting to try to figure out why, although a few reasons do come to mind including fussiness/costs for the robotics. That's part of where industrial automation/robotics is where the $ is I suspect.
I refuse to program PLC tho
Would you do it for $200/hour? ;)
@@ChangeNode ye... I would, but I would have to relearn a lot. It's a whole different way of thinking. Different design patterns. Parallel. It's not an easy transition.
@@ChangeNode I would not be comfortable taking that amount at my current skill level. That's for sure.
@@test5095 oh, sure, it's more of just a comment on PCL. :)
@@ChangeNode yeah sure, but I don't know if it makes sense because 88,000 seems to be the average salary. Even in the us
The industry isn't structured for the multi disciplinary world of robots. They want specialists, not generalists but building robots requires a lot of generalists. Still waiting for blockchain for robots.
Hello I am a junior in college with no internships. Should I keep learning spring or jump to rust for now and the future?
There are way more jobs right now for Java/Spring I think. Longer term you are going to want to get very good at being able to pick up languages fast. Rust and Java are both C/C++ descendents so that part isn't too bad. The tricky thing about a new stack is always the libraries/SDKs. FWIW one of my favorite books amzn.to/4cbUYts even if you don't go through the exercises it's a great read.
I'd get a part time job/internship for this summer lined up ASAP. My guess is market will still be soft in 1-2 years but (crossed fingers) will be better after that. Your #1 is to get ~3 years professional experience on your resume.
@@ChangeNode does intern count as professional experience?
I actually think the term AI/Robots and similar are used more for marketing hype. Automation is a great tool that makes sense for business, but getting eyes on newspapers and TH-cam videos. For that, you need something more exciting. :)
atleast i can say that i was 4Y ahead for getting in robotics .
The uncanny valley is very much a spectrum and you must be on the more sensitive end of it cause I don't find these robotics movements uncanny at all.
Oh, depends on the robot. But absolutely different people have very reactions. Same for stuff like motion sickness - eg there are a few apps on my Quest 3 that my son is fine with that make me instantly queasy.
My guess is that they a) they will tweak to fit for purpose and/or b) we may just get used to 'em.
At 14:20, if I design a robot, I'm going to use Rust instead of C. C is a minefield and should be avoided.
Yeah, Rust in the Linux kernel was huge. Under the covers so much stuff still relies on bindings to C functions, so still worth learning at least how to read IMHO.
I would need to have a degree to have a job in robotics?
I think that mostly depends on the role/field, but generically yes. I think you could get a long way doing self-study but the sense I get looking at the job descriptions on Indeed/LinkedIn is that there is just so much to cover I think it would be very hard to break in without a degree.
You might want to check out the open source robotics stuff I mentioned in the vid and just get a feel for it yourself.
i was thinking about getting a CS degree but after the latest news i changed my mind and now Iam looking for safest engineering majors from automation and Iam interested in two (mechanical engineering and electrical engineering) so what do you advise me and are those really safer than CS ??
CS degrees are still good. You may hear a lot of doom about how AI is going to replace all software engineers, but most of that is nonsense. The nature of software engineering is definitely going to change as we learn to adapt with the tools, but you will still need software engineers. If AI really does replace software engineers, then every type of engineer is toast lol.
If you go to a good school, have a great portfolio, and get good internship experience, you will have nothing to worry about (the current job market is tough, but things will get better within a few years as the market recovers). CS overall is still a very in-demand, fast-growing field leading the frontier of new technologies like robotics, ML/AI, etc. If you don't end up majoring in CS, at the very least learn to program and keep up to date with ML and that will give you an edge in whatever field you end up going into. Good luck!
@@pengtroll6247 thank you so much ❤️
Accounting is always hiring. If you want safe engineering, civil engineering.
Electrical engineering is good as AI still needs hardware. Also there’s enough code involved to pivot to developer role fairly easily. If I were doing a degree today I’d take electrical eng.
Sigh. First, if you’ve been using LLMs to write code for you and you’ve concluded that software development is over? It tells me you’re either not a very good developer, or you’ve not used LLMs very much for code. OpenAI, Llama2, Claude. They all make mistakes. A lot of mistakes. They will always make mistakes, because LLMs are NOT REAL AI.
This will be a productivity tool for you to use in the future. If companies hired developers to write code for them that you can copy off of StackOverflow. Guess what? They’re going to pay you to copy code from an LLM and make it work.
You guys all sound like Boomers in the late 90s when they thought computers were going to take all their jobs. Here we are 25+ years later……
In general, time for some folks to get back to actual computer science and doing some hardware stuff
Me with a robotics degree who pivoted to softdev: welp, here I go back
(but not really)
lol yeah. tbh part of the point of the video was to kind of explain just how hard it is. Don't want to discourage anyone but also wanted to explain that it's not actually likely to be easier.
Pondering doing a similar one on plumbing, ha
Liked and Subscribed.
you got a new subscriber will
If you've been doing software development for "several decades" you would have came out of the womb with full knowledge of cobol.
lol
started coding when I was 10, BASIC on Commodore PET and Apple //c
first paid job was when I was 16, writing macOS stuff in Pascal
macOS Pascal + a little bit of 68k assembly in college, working for a lab converting bits of COBOL, FORTRAN, and BASIC to "modern" Pascal.
Been going ever since. Apple & Symantec in the 90s. Startup from 1999-2001. Consulting from ~2002-2016...
Anyhoo... yeah, paid since 1992, FTE bonafide SV tech since 1995. So, yeah, that's decades. It's completely insane looking back on it...
If you didn't have access to Commodore PET in your mother's womb, then your parents didn't love you.
If we really think about it, with the industry really care if AI robots working in our homes put the cat in the dishwasher? I mean, we have airplanes being flown by human beings falling apart mid air, and nothing is being done about it. You have people who built those planes on camera, saying that they would not fly in those planes.
If Robotics explodes- then so will the proliferation of abstraction/frameworks to interface with them. Very few people will actually be Robotics/Systems Engineers, in fact they will be 2 or 3 levels abstracted away because protection of the source code will be of upmost importance. I think it will create newer DSLs and scripting languages for people to interface with. It will be the same as it ever was. Instead of a funky app on a smart phone ,it will be a funky app on a smart phone to interface with a robot in your house like a roomba.
Definitely disagree with companies being tone deaf to terminology around robotics and ai. None of these companies care about what the public thinks. They want to induce anxiety in investors and they are good at it. If the work a company is doing seems scary and paradigm changing, you better own a piece of it.
Yeah, that tracks. The humanoid form impressing the investors and the rest doesn't matter. Hmm.
We’re going to build our own masters!
He said that you only save about $5000 per car, but that is wrong. Image if robots are used all the way down the line. You have robots harvesting the raw materials, You have robots creating the steel, You have robots building the robots to do all these task. Soon the cars cost is under $1000. With little to no humans any where along the line, that is the future.
Yeah, there's absolutely a line to that eventually. In another thread (on the what is AGI video?) it meandered into a way to measure the capabilities of a robot in terms of how close it can get to full self-construction. Everything from resource extraction, processing, and assembly. Gets murky eg when considering just using existing things like chip fabs vs actually being able to do chip fab.
+1 for it being the future, but the timeline is murky...
I am worried that people are looking at this with a cloud of misconception, they think their job is safe until those humanoid robots become common
The WFH during the pandemic as I see it now is a double edge sword, to all jobs that could function without being in the office could potentially be replaced by AI in common computers not robots
Been looking at this for months, I conclude there will be job reductions in the future and it will be massive, as early as now I am trying to prepare for this financially, physically and emotionally
I hope people, in just small steps daily, start preparing for this
The next video on AI Robots cost is more inline with some of your concerns I think. This stuff is moving very fast and it's very hard to keep track.
WRT preparing... yeah. It's a very strange time.
greed is the issue , it always have been, you need to regulate greed first
You deleted my comment on your last video. Not sure what I said wrong...
I haven’t deleted any comments…? Weird.
Maybe you triggered some TH-cam auto mod thing?
@@ChangeNodeI don't know. What I said was over all I agreed with you that tech is more than just software engineering. I shared my experience with the 2000 doot coom bubble and I said I was umemployed for so long I ended up joining the Army out of desperation.
I just thought you (or someone else) thought I was being pro-military and trigger the comment delete...
@@bigbigdog no idea, wasn't me. FWIW I did check and there were two comments held by TH-cam auto mod for review, but they weren't yours and I just approved both of them. AFAIK you can try resubmitting and see what happens. Did TH-cam send you an email or anything? If not my guess it that it just got eaten by the server or some connectivity thing...?
@@ChangeNodehm... Thanks for checking. TH-cam has been acting weird for me lately.
It's ok. The other comment was pretty long and I kinda forgot what I worte anyways. tks.
I don't think it's a good idea. First, only very niche hardware spaces are still okay: FPGA, ASIC... but forget about GPU, CPU, and above all MCUs. These are massively oversupplied fields with thousands of applicants only from India, Pakistan, Indonesia, Philippines etc.
Second, most software developers need years to become proficient on hw.
Yeah, most of the robotics/hw folks responding in the comments have all basically said "don't"
Part of what I wanted to do was gently explain to folks how much work it is. Kind of like being a doctor or electrician, I think a lot of folks don't realize how hard/difficult it is, but if you love it you love it...
Hmmm might be time to look at Rust
Hahaha I also use Siri pretty much for set timers while cooking 😂
Hey ChangeNode - CourseCareers sent you an email about a paid partnership. Let me know what you think.
I would like to see japanese style vending machines and robotic restaurants. for 2 big reasons.
I don't about where you live, by my experience with fast food is: it's overpriced and the people who make it are NASTY.!!!!!
because of that, i no longer eat at restaurants.
Funny how you mention the investments going into 1X/Figure but leave Tesla's Optimus out lol.. Mass manufacturing is what matters and Tesla will own on that front
Tesla? th-cam.com/video/UTZ6Lvh5gL4/w-d-xo.html
3D printing a tooth is a solution looking for a problem. We have super easy ways to handle that without trying to use a 3D printer.
My dentist had what was technically a CNC system that would cut a tooth or cap 15-30 minutes based on a scan. He'd then order out for the full version but it could take days. I don't think it was that expensive and it sure made life a lot better as an interim solution. FWIW.
Your math on replacing labor might save 10% doesn’t account for the humans replaced in the supply chain, so it’s flawed math. You need to look at all the materials used to build cars or the items on shelves in retail. It’s not just the retail staff that matters. It’s the savings in the entire supply chain.
Oh, absolutely agree. There's a bootstrap test for robotics - how much of the entire supply chain can you automate with a robot?
Part of what I am focusing on right now is the transition period and minimum economic viability of this stuff. Once it starts getting any deployment I think it starts becoming a self-reinforcing loop...
@@ChangeNode Appreciate the reply. 👍
Nah, better build out cognitive assistants. There is no shortage of physical workers, but cognitive workers are very rare. We got a lot of actors impersonating cognitive workers, but that’s quickly going away.
First I've seen term cognitive assistant. I think I know what you mean but I'd rather ask - more details on what you mean?
@@ChangeNode He's too stoic to elaborate further. Understand, or be left behind.
I'm sure you've seen this demo by now?
th-cam.com/video/Sq1QZB5baNw/w-d-xo.html
How might a developer pivot into robotics without ready access to great hardware?
RE figure, between that and Devin dropping essentially right after my vid looks like I've got another vid to do lol
RE: robotics a lot of the controllers/subcontrollers look like the are driven by tiny little Linux hardware, you can get pretty far with C/C++/Rust on a comparative potato. Local LLM stuff is a lot more about the big box, although I have played with a few sub 2gb LLMs that could run on my iPhone. Check out the open source robotics page and you could find some pretty inexpensive stuff to play around with.
@@ChangeNode Thanks for the helpful response. Really appreciate it. And yes, by all means, please post a follow up video. Would love to hear what you have to say. Subscribed and staying tuned...
I prefer no zip hoodies....
Here in Seattle it's flannels 😂
Oh so weird. That comment was supposed to be on the video I watched before THIS one in the queue...
Energy is the answer, with cheap energy more things are possible including make water/food from the air. AI/automation needs lots of energy, solar/wind/batteries are not the solution.
I don't even know anymore
www.reuters.com/technology/microsoft-buy-power-nuclear-fusion-company-helion-2023-05-10/
www.theverge.com/2023/9/26/23889956/microsoft-next-generation-nuclear-energy-smr-job-hiring
It's getting awfully cyberpunk out there
@@ChangeNode Buy it and lock it away or use it to power github and use everyone's code to improve their AI?
Robotics sucks. I did my masters in robotics, and tbh the pay in that field is 1/4 what it is other places. Meaning, specifically, similar roles pay literally 1/4th as much. I could not justify staying in the industry so I changed to Web 3 and edge computing (solving similar problems) and I now make 4x the salary.
Web3? What part?
How are you feeling about job security/finding another job if you needed too...?
@@ChangeNode EVM work. Its quite easy for me to find work at least, and I do get messages from recruiters 2-3 times a month. Cant predict the future tho!
Instead of everyone going back to study robotics, maybe robots start fixing themselves.
Would you agree that it's a good thing that Earths population is on the decline, just as we enter the age of AGI powered androids and a decreasing need for workers?
RE: Earth population & demographics, that's a big topic and it starts to get into a lot of cultural, political and economic complexity.
So... I'm going to punt on this one mostly for now. I guess I will say that if a country / system can't manage to maintain population levels, eventually it will disappear, which seems less than ideal.
Most of the objections to growing population seem to be rooted in climate and resource impact. I personally do think that there's no reason we can't dramatically reduce the resource impact for individuals assuming there is system wide support. If we can't figure that one out... well, I guess that in the long term that's yet another answer to the Fermi paradox. Sigh.
FWIW en.wikipedia.org/wiki/Malthusianism
Alex Jones in a parallel universe
Your comment that a Restaurant becomes a glorified vending machine, so why would anyone want to go is missing the point of restaurants. We don’t go for the servers. We go, so we can have a nice prepared meal we don’t have to cook and we don’t have to clean up after. It’s not for the human servers.
Yeah, you can break down the components of going out in several ways. Human contact? Ambience? Don't have to cook?
My point was more that for a certain class of restaurant, just having the experience turn into a vending machine was missing the point, whichever of those components is most important. I have seen videos of sushi restaurants in Japan that are basically entirely served via order screens + a conveyor belt. I think someone still comes in to clean those up by hand.
Either way, my guess is that the human contact is the last thing to go. If you had a figure.ai robot that would prep/cook/clean up then why go out? My take is human contact.
@@ChangeNode I guess it depends on the type of restaurant. Fast food or Starbucks etc is primarily cost, convenience and taste. I don’t care if I’m served by a robot if the other 3 are good enough
If it’s a Chop House looking for a great dinner it’s about food and ambiance. Convenience doesn’t matter much. Cost is what it is…. A great server is a nice to have, but often their value add is getting our orders right and quickly. If a robot “could” do both of those things and at a lower cost I wouldn’t complain. 😊