Chapters (Powered by ChapterMe) - 00:00 - Coming Up 01:01 - Jared is fired up about vertical AI agents 07:25 - The parallels between early SaaS and LLM’s 09:09 - Why didn’t the big companies go into B2B SaaS? 12:25 - How employee counts might change 16:25 - The argument for more vertical AI unicorns 21:31 - Current examples of companies/uses 35:22 - AI voice calling companies 40:04 - What is the right vertical for you as a founder? 41:36 - Outro
I loved the point by Jared about 3 major types of internet companies: 1. products/services that people obviously need: mail, search, maps, messengers, ... - dominated by the largest corps 2. not obvious consumer products - AirBnb, Doordash, Uber, Instacart, ... - born from startups that were more agile and creative 3. B2B SaaS - multiple verticals, not dominated by large corps it kinda creates a picture for me
3 things to vertically differentiate 1. Vertical agent has exclusive access to high value dataset . Ex Corrosion detector agent has exclusive access to worlds largest refinery corrosion dataset 2. Vertical agent owns a key decision making node in workflow. Example : Next best action decision in equipment troubleshooting buddy sgent 3. Vertical agents own end to end workflow. Ex : Regulatory reporting agent in Pharma Trust that helps
there will end up being standardization of these agents and thus drive agglomeration of companies when no one has any intellectual property worth keeping any one company separate from the rest
It's wild to see how often you, the Top Startup Advisors, are shifting your perspective and advice lately-it’s a real testament to just how fast and profoundly things have been changing these past couple of years. Feels like the ground under everything is constantly shifting!
@ it’s a version of the optimal stopping problem. It’s not really known. Anyway, using Lean startup techniques should avoid just chasing trends like this. Somewhat.
I don't follow the narrative that these YC folks are pushing about boring jobs. If people working those jobs could get something else they would. Making a case for AI removing boring jobs just so it "allows" the current employees to leave is just nonsensical
100%. We‘ll have a massive issue with this. The majority of people just want a normal job to pay for rent and living. Only a few are qualified and capable to reskill to deal with AI to stay ahead of the curve. AI will make intellectually challenging jobs redundant as well, and not because they are boring but because it can and will be much more time and cost efficient, which is and will be the guiding economic principle in a capitalist society.
Life is more complicated than that, I used to do data entry but I lost that job 20 years ago when it was outsourced to India. So I went back to school and I got a degree in software engineering. After a few years developing software, nowadays I'm doing finance and paperwork mostly simply because it's easier and I get paid the same somehow. Both will probably be automated somehow at some point by AI. Then I'll just go back to school to find something else. Don't forget work is basically doing something someone else does not want to do. There will always be something someone does not want to do. At least until we have robots doing everything.
@jean-phil 100% . Life is more complex which is why it can't be summarised to AI will "allow" people to not do boring jobs. Really nice journey. I love your grit
Some great products have come through the Y Combinator pipeline - but this conversation feels extremely cold and out of touch to me. It’s hard to get excited about the prospect of AI taking away the jobs of entire teams.. feels very out of touch
The only aspect I find exciting about AI if it is used to augment my teams jobs. If people start losing their jobs en mass then this idea needs to end.
Only in a world where all workers own a part of the benefit after they’re replaced would this be a good idea. If workers start being replaced for the benefit of a few there will be major revolts.
The way creative destruction works doesn’t care about how we feel. 90% of the worlds employment was in food production once, it isn’t any more and there are many times more jobs. Fearing the future can’t change it that much.
Light Cone Podcast - Vertical AI Agents Discussion Timestamps Introduction & Historical Context (0:00-4:04) - 0:00 - Discussion of progression in AI capabilities over 3-month periods - 0:05 - Introduction to vertical AI agents replacing enterprise teams - 0:17 - Evolution from OpenAI dominance to competitive landscape - 0:44 - Introduction of hosts Gary, Jared Harge, and Diana - 1:04 - Jared's enthusiasm about vertical AI potential - 1:28 - Prediction of $300B+ companies in vertical AI category SaaS Industry Analysis (4:04-9:09) - 4:04 - Historical context of SaaS evolution - 4:44 - XML HTTP request as catalyst for SaaS boom - 5:41 - Three categories of successful SaaS companies: 1. Obviously good mass consumer products (won by incumbents) 2. Non-obvious mass consumer ideas (won by startups) 3. B2B SaaS companies (300+ successful companies) - 7:53 - Discussion of why incumbents didn't enter certain markets - 8:47 - Analysis of Uber's early regulatory risks B2B SaaS Deep Dive (9:09-12:28) - 9:09 - Analysis of why incumbents didn't dominate B2B SaaS - 10:34 - Discussion of bundling vs. unbundling in software - 11:35 - Enterprise software pricing models - 11:51 - Problems with traditional enterprise software Employee Count Changes (12:28-16:25) - 12:28 - How AI might change company employee scaling - 13:02 - Current unicorn employee counts (500-2000 employees) - 13:31 - Shifting hiring priorities with AI - 14:43 - Historical perspective on company building - 15:45 - Example of engineering approach to marketing Vertical AI Examples & Analysis (16:25-24:22) - 16:25 - Case for 300 vertical AI unicorns - 17:01 - Discussion of enterprise uncertainty about AI needs - 18:30 - Consumer vs. enterprise adoption patterns - 19:50 - Why vertical AI could be bigger than SaaS - 21:37 - Real examples of vertical AI companies - 23:53 - Case study of QA testing automation Customer Support & Specialization (24:22-29:01) - 24:22 - Analysis of AI in customer support - 25:51 - Example of developer support automation - 26:30 - Discussion of Power Help case study - 27:42 - Importance of vertical specialization - 28:38 - Why specialized solutions win over general ones Management & Scale (29:01-35:22) - 29:01 - Discussion of Coase's theory of the firm - 30:02 - Rippling case study - 31:06 - AI's impact on Dunbar's number - 32:16 - Example of AI-powered employee communication - 33:43 - Rippling's organizational structure Voice AI & Industry Evolution (35:22-40:04) - 35:22 - Discussion of AI voice companies - 36:17 - Case study of Salient in auto lending - 37:31 - Evolution of voice AI capabilities - 38:11 - Historical perspective on LLM applications - 39:23 - Rate of AI progress analysis Closing Thoughts (40:04-41:53) - 40:04 - Advice for founders on choosing verticals - 40:19 - Importance of finding boring, repetitive admin work - 41:34 - Final analogies to robotics - 41:47 - Closing remarks about targeting "butter passing jobs"
Yeah, but what is preventing SaaS companies from just pivoting their tech stack or integrating vAIs into their offering? Like trying to sell into these categories is extremely hard, even if the tech is cheaper, faster, better. These customers are exhausted.
Back in the final semester of senior year of college, I remember being excited to tell my career professor that I had a job offer. Initially, he seemed delighted, and asked, "what is the job offer?" "Sales", I replied. "Not my first choice, but base pay actually seems to pay rent." "Sales? Sales?!" He nearly collapsed before bursting into uncontrollable crazed laughter. "EVERYTHING IS SALES! EVERYTHING! EVERYTHING!!!" His laughter devolved into a choking wheeze, forcing him to pause and hyperventilate for a few minutes. Tears weeping from his wrinkled eyelids, he reaffirmed in a quieter voice, "everything is sales... everything... everything..." In that moment; I realized what I was in for, the dark reality of the working world.
@@scott_strang The innovators dilemma is top of mind and commonplace at most orgs today. Although how previously relevant this concept has been, SaaS is a hyper aware segment to this notion since they previously implemented this exact playbook.
Great point. These traditional SaaS will just buy these emerging vAIs and plug it with their existing customers. They own the last mile with customers, and Enterprise rip and replace software is not an easy switch.
I guess there is some confusion here. SaaS is a business model that won‘t go anywhere. On the other side, Agents is a tech progression of the AI advancements. These are two different things. What will be the challenge for SaaS companies is to adopt their value proposition at a higher pace to avoid commoditization.
Loved this conversation! Two takeaways that stand out for me are: 1. The focus on founder-led teams at Rippling suggests that domain expertise and entrepreneurial thinking remain crucial even in an AI-driven future. Collaboration, Critical Thinking, Creativity and Communication will still be the foundation of how humanity progresses regardless of the tool. 2. The most successful AI startups might not be those with the best technology, but those with the deepest understanding of boring contextually-specific workflows. Two questions: 1. What does these 'new rules' look like decentralized? 2.With World Coin on a mission to bring a billion+ new 'players' into the game, how might we empower these traditionally underserved global communities? What a time to be alive!
@@ispinola It's also what I have been thinking to be honest, because what ends up happening is you layoff so many high income earners, lowering the tax base. In return the Tech giants make their money back (maybe) but in the end no one will have the money to spend if there is no job. Basically selling to businesses, banks where most customers won't even use the features
Umm, I'm going to need AI model builders to take a vacation. Come on. Just let us get rich off our apps before you outdate us all a week after launching. Y'all have no chill.
The lump of labour "fallacy" is one that the tech community loves. "Don't worry about job automation today because these are simply yesterdays jobs. Embrace tomorrows jobs today" they say. Except, no one stops to consider any potential retooling time for the economy. It might well be that there will be new jobs "tomorrow". It's just that today there are millions of people waiting for that tomorrow- and it isn't coming anytime soon. What to do in the interim, hmmm? When will people wake up to see this I wonder ?
Also all we have is intelligence and physical labor, new jobs created by industrial revolutions in the past still needed both. But AI and Robots replace both. What do we have left? Creativity? Emotional intelligence? AI is doing pretty good on those fronts I think there will still be jobs, but just for the most adaptable, resilient, and intelligent people
True Ai agents will transform businesses to next level. I am building AI agents for E-commerce, Real estate, Tourism and Healthcare companies. I increased their productivity from a 12 member team to 35 members team without changing the team size. Now in just 25 days of deployment they got 3.5x more revenue from last months. Results are motivational.
🎯 Key points for quick navigation: 00:55 *🚀 Vertical AI agents are set to revolutionize enterprise functions and replace entire teams, marking a significant paradigm shift.* 01:09 *📈 Vertical AI agents could spawn $300 billion companies, paralleling the impact of SaaS in transforming industries.* 01:53 *🌐 Many startup founders underestimate the potential size and impact of vertical AI agents, similar to early misconceptions about SaaS.* 03:15 *🖥️ The evolution of XML HTTP request in 2004 catalyzed the SaaS boom, enabling richer web applications like Google Maps and Gmail.* 04:25 *🎯 Just as SaaS disrupted traditional software, vertical AI agents represent a new computing paradigm with vast untapped potential.* 07:13 *🛡️ Incumbents often miss disruptive opportunities due to the risks involved, akin to early reactions to Uber and Airbnb's regulatory challenges.* Made with HARPA AI
I started coding before AJAX, that's not what caused SaaS to happen. It was one of many improvements. The main thing was just the web and dynamic scripting languages which enabled it.
Yeah I was thinking that. Also, they talked about QA as in testers, but there are many orgs that are much more efficient in their QA efforts and would not swap an human for an agent, not until "AGI". It is sad to see YC pushing the Altman agenda.
Vertical means, serving a industry from A-Z with your SaaS. you could choose any industry and build suite of tools for them to do everything they do. i.e: if you take milk manufacturing industry, your vertical SaaS would cover everything this manufacturing industry needs.... from getting the milk, storing, paying, selling, accouting every aspect is done by your suite of SaaS. This is why its called Vertical. You choose a industry, you go down deeeper and deeper
It's a term not a catch phrase. It has a definition. I'm a vertical construction layout instructor. Meaning primarily MEP ( mechanical, electrical and plumbing) so we segregate civil, industrial, and vertical. ( Residential is not within my scope) Each industry utilizes cross industry definitions.
CEOtech... they touched on this but glazed over it - but one of my biggest takeaways. We're working towards a model where insights and analytics and answers don't have to pass through seventeen layers of management to get to the decision makers... to the CEO or Founder or business owner. Even skipping over the CFOs and COOs and management. THIS is the big play.
Essentially leaving a single developer who will act as a project manager. No more junior, business analyst, data engineer, data scientist, scrum masters.
Hunting for the boring repetitive jobs should be every AI startups focus. A friend and are building a basic MSP and have locked on with a client who employs many people to perform basic repetitive services. We get to look at this clients work efforts through our AI lenses and the number of opportunities we have categorized is a bit amazing. When time permits we are building out our flavor of an industry vertical agentic solution that we feel applies to multiple vertical service types of SMBs.
At my company, we have an IT department that is protective of our SQL & cautious of any Apps/Scripts developed by employees who ACTUALLY have ideas on what can be done on local process levels. AI in a company like mine (Larger lumber/truss/panel company) Multiple locations/States etc etc. we need designers/production managers/inventory specialits to be using Claude/OpenAI etc and building now, but actually getting that deployed IS THE major holdup. Any advice on how to get this pushed forward. I imagine my company is 95% of the companies OUTSIDE silicon valley.
Also BTW we use complex software that will take years to redesign as AI software, but an AI plugin- that help opperate that software like Revit or Autocad, is a much more helpful system in the near term. Why doesn't Mitek/Autocad & Revit have built in AI assistants rolled out already? How are they companies not deploying these tools within weeks?
The future of work is that humans do not have to work anymore because AI/robots will be able to do all the work for us. We‘ll just have to come up with the societal change to actually make that work.
Correction, Information Unlimited, then TeamScreen Solutions was a SAAS for the employment background check industry starting in 1997 and yes with XLM before Salesforce, etc…..
21:05 "do random data entry or approvals or click the software" .. Yes and No ... the way I see it, it's more complicated than that. If you take "approvals" for example, that's not going to change, it's not like we are going to let AI approve anything. While AI might help analysing what needs to be approved and will make it more efficient, it's not going to remove that step. Regarding "random data entry", even if AI gets really smart, the reality is that Enterprises often have a lot of "legacy" systems and sure there is probably too much time spent on copying data between systems manually but if it's not something we have fixed via APIs yet, what makes you think AI will solve this ? ;p I mean, I like using chatbots to be more efficient and I'm sure there will be a lot of opportunities for vertical AIs but I think a lot of bureaucracy will remain. At least for the next 10 years, until Enterprises slowly switch to more AI friendly software/tools. You need to remember, a lot of software is custom and a lot of enterprises are quite niche, it can be hard for an AI to make sense of these. Sure Casetext appears successful but they "automated" a job (lawyer) that is very dependant on reading documents and with over 1 million lawyers in the States, it's a very common job somehow. I could see the same thing happening easily with programming for the same reasons. And sure there are other "quick" wins probably like the video game and film industry because both deal with digital information and sell digital products. But for the rest, not so sure. For example, I think these will take a longer time to be "profitable" for AI : retail, government, manufacturing and unpopular opinion : education. Finally, I don't know enough about finance and insurance to tell for these industries. Ok one last : "administrative work" .. it depends on what you think is "administrative work". Sure "scheduling tasks" and "organizational tasks" are easy wins. And maybe one "admin assistant" will be so efficient that, as a result, he/she will be able to work for multiple managers at once but I don't think it that it will replace them completely because some tasks can't be automated easily. Or maybe some will fire all admin assistants let the managers do the remaining work .. but that work will remain and more managers might be required as a result. Like you mention automating recruiting, sure it can automate some tasks .. but the whole job ? Again, I think it can be a great tool and will make people more efficient but will not replace entire teams. I mean, it will happen here and there but it will not have a massive impact on the job market until at least 5-10 years.
I appreciate your comments around how Enterprise software is not best in class when compared to some of the newer offerings, especially emerging tech from the LLMs. I’m actually keeping track of the tech adoption lifecycle and how this could change who wins and who loses in 2025 and beyond.
Would be nice if you made a small list in the description (or a written post somewhere) of all the companies/tools/products you talk about. Then I can easily find them all. Sometimes I try to type in the name of a product you mention on google but I can't find them - unfortunate for them! Surely some AI tool could do this for you without much effort :)
being a person in the oil and gas industry - the problem is we believe that AI enabled apps would automatically share our confidential data with the rest of the world. Could you do a series on privacy of data if we use these LLM infused RAG data within our organisation...
with all these LLM agents running companies, who is gonna buy the products?? This is my constant question. 80% of jobs are automated away, who is still buying stuff?
100%. The transformation will be insane! But also there'll be many more vertical AI agents than there are SaaS now. Current B2B/B2C soft cos are not super niche down bc they were difficult to build, but AI democratizes software the same way online manufacturer marketplaces democratizes e-commerce. We'll have Vertical AI Agents (companies) helping executives or highly driven people with every single task in every single niche.
@@AIJay-Pro Yeah 100%. Maybe a bit lower price than SaaSs, but you'd still have to pay for intelligence APIs. With software there's no marginal cost of replication, but with Agents you have to add Intelligence buying to the COGS (cost of goods sold). What do you think?
The trick is going to be business moats - how to differentiate from every other company that is now quickly capable of doing the same thing. Feels like a land grab and the winners will be more those that nail the marketing and branding. A few will develop unique tech inside the stack - but a lot will be "wrappers". Kind of like all the car companies that surfaced initially and now we're down to a handful.
The one thing AI hasn’t been integrated with is our financial world. As a digitally native agent the default rails for executing on this financial vector is crypto. AI agents in crypto are taking off like GOAT, ai16z, aixbt, Freysa AI. All are examples of these vertical agents I wish was mentioned lol
Anyone know how you can build an AI agent to help with marketing a product or maybe help to read invoices and receipts created by different companies and transform it into Excel data and generate charts.
Agents won’t remove software layer completely. They will merge, and software will be the orchestrator, because we need predictable outputs, which AI can’t produce reliably enough.
Two amazing thing side note things said that were deep cuts - It more of a butter passing job, it kind of sucks. This is from Rick and Morty. - Open aperture of the context window of how much information they parse. This would increase their Dunbar's Number but was just well said!
Selling a LLM vertical - construction , typically a lagging industry but with everyone using chat gpt it has bridged the gap in people “not being ready” for ai vs - yes I need this in my buisness
Nah, we already have a platform to build multiple use cases in prod that covers many verticals such as clinical, marketing, & agriculture. Including a few others I cannot list due to NDAs. We take a totally diff mathematical approach-experience matters.
Can anyone really understand what this lady is saying? At 25:51, she mentions a dev tool company called capillary ai (or something), but Capillary is not a dev tools company. Even the Transcript on right is not clear ... Such a poor conversation!
Oh man any Y Store users out there? With the GUI prog lang where you click a button to insert command, then enter command instructs. Templates were LISP-y and quite flexible though
I find a lot of this discussion non-sensical and without substance. If you end up replacing humans with tech at the pace at which they are suggesting - what will the businesses do? Businesses exist to serve human needs - whether its B2B or B2C - its still human needs. So, if humans aren't part of the businesses - then would these techs serve other techs?
Undeniably we are living the start of the next World Wide Web boom. Currently LLM's are the big talk in tech, but even right now they are just a new born baby, all the big LLM companies are basically good at 1 thing, written language, as it only understands text. I think whichever company will get the Large Visual Model (LVM) right, will get a huge start in this AI boom
29:29 isn't the actual limit scaling equity and compensation? i mean, they just talked about how demotivated were the high level management at google to go after Uber, but if they DID go founder mode with G resources they would crush it.
AI vertical agent? means u don't need to get bunch of butter giver herd instead of count of little that fulfill your visons without efforts to extract butter with 100 hand. so it increase your human capacity as individuals.
4:00 Er, didn't we used to call this action a Remote Procedure Call? It wasn't easy because there was so much comms and security baggage. SOAP made it a bit simpler. It all depends how you package it.
These people are awfully calm about removing jobs in a society that requires them to survive, and provides little-to-no support for humans who are caught in the middle.
None of them care. Or, at the very least, they just assume that the problem with just sort itself out with everyone becoming entrepreneurs with AI workers
Yeah, I mean the reality is that if they don't do it, someone else will. There is no stopping "progress" so you might as well ride the train so you don't get left behind. But you could at least show a tiny bit of concern for the people in the departments that will be shredded down instead of only lamenting how you can't be completely honest with some of the people you're selling to because their jobs are threatened. There's no question that from a startup's perspective, AI is super empowering and you don't need to be bogged down by hiring as often as you grow. You can focus more on the work and innovation, but remember that you'll be selling a lot to established companies that already have employees and you're helping those companies trim fat - but that process is also putting people out on the street. At least acknowledge that truth. This will be both an incredible and very difficult time for humanity.
To respond to the comments about AI replacing staff: This is in fact the ultimate test of whether or not your AI passes the ACID test. If it is just helping or assisting, it's not there yet. If you suddenly don't need a staff member, now you are getting somewhere.
I wonder how concerned the VC world is about Salesforce and Servicenow eating the agent marketplace alive serving enterprises and SMEs which in turn use the agents crafted thru SF and SN to crush smaller businesses.
If these AI Vertical agents are going to replace all the employees, whom are they going to see the products too? People don't have any income to buy anything.
31:08 Dunbar's number: "there's a limit to how many people you can have meaningful relationship with" I read about it. Do you mean we would be able to extend the limit or we won't have to deal with so many people because many of them will be replaced with AI?
I think you'll still need people to build teams, of real people, but what will change is where the team coordinator requires a greater understanding of the tools available, as well as a priority to automate tasks. There is a danger of the automated stacks rotting away over a period of time, with vulnerabilities not being noticed until it is too late. So security overview has to be a key part to automation. Unfortunately, security is generally a lower weight problem to startups.
Hi, I’m a multi-billionaire who’s looking to break a trillion but I’m tired of listening to elevator pitches from founders and CEOs. What completely vertically integrated (human-free) AI systems can you recommend?
I likI like what Jared said about the Engineering mindset being applied to other disciplines like marketing. As an Engineer, I always felt that I needed a "marketing guy" to sell the service. But in fact, this is not true, it just requires the same mindset to a different and messier field.
Microsoft knew about Ajax before Google, and wanted to create an outlook online. However felt these super responsive web apps could impact their windows business
One of the reasons the incumbents didnt get into B2B SaaS as most of them are getting into the infrastructure space - storage, compute and now chips for AI.
I like how you guys talk of companies with only 10 employees. What will the unemployed ones do? Who will consume products made from these LLMs, and where will you guys get funding to fund other startups if we hand everything to these machines?
If so many people are at the risk of getting replaced by AI, I wonder who the consumers will be? How can anyone afford to buy the products/services and therefore how can these companies turn into billion dollar businesses? After all, not everyone is likely to become an AI expert, right? Of course, there could be some areas which are less/not impacted by AI as much but since the effect of AI is expected in so many industries, will this not cause massive unemployment and therefore less people who can actually afford to buy any service whether AI or non-AI?
that's what these lunatics dont understand They believe govt will hand out UBI to everyone These people keep on talking to themselves in their echo chambers AI bubble will burst, most of these startups will come crashing down
I feel like a new AI startup got their wings every time they said unicorn. Was waiting for them to say they've worked with hundreds of Fortune 100 companies. Can we get back to a unicorn meaning a once in an interval event? If you are finding them all the time, they are not unicorns. Could be great companies, industry leaders, but not a unicorn...
@javierserrano4695 was the case when it wasn't normal. The term unicorn infers rarity, which will change over time as the norm for valuation shifts. This should not be used so loosely, as with this crew, or it (and they) loses all credibility. You start to assume all things out of their mouths are hyperbolic nonsense (insert a clip of dirty used car salesmen) vs. credible and measured statements from experts in the field.
the playing field is open.. and as Gary says the consumers will have choice. It's great that there won't be a monopoly on AI... or would we have the AI giants like how we have the tech giants??
No Microsoft basically has a monopoly on AI. They are just hiding it a little bit. They own basically all the infrastructure that AI runs on and they have their fingers in many of the big AI startups so their monopoly is not going anywhere
so every company is just now selling AI agents who can do anything? So all companies are selling the ssme thing? The further you follow this rabbit hole the more you realize none of this makes any sense.
They are missing the key point that there are 300 incumbent saas companies sitting on treasure troves of data and workflow knowledge that will implement agents for their locked in customers. Incumbents will capture most value here.
"Vertical ai" is made for specific industries and use cases. Addresses specific challenges, automating a workflow "Horizontal ai" is made for a broad use case across different fields/industries.
There should be a universal web based platform & protocol for development and monetization of AI agents following the immensely successful architecture of the DNA/gene agents. Working on a prototype. Connect if it sounds interesting. It should be the base for continuous innovation and re-skilling of billions of professionals.
@@imkailin From an evolutionary perspective, we can consider the genes as successful natural intelligence agents evolved on DNA engines in the past billion years. We should replicate this successful DNA engine architecture from Bacteria & Eukarya into an AI Internet protocol that should facilitate continuous AI skills and services monetization for the AI economy.
Chapters (Powered by ChapterMe) -
00:00 - Coming Up
01:01 - Jared is fired up about vertical AI agents
07:25 - The parallels between early SaaS and LLM’s
09:09 - Why didn’t the big companies go into B2B SaaS?
12:25 - How employee counts might change
16:25 - The argument for more vertical AI unicorns
21:31 - Current examples of companies/uses
35:22 - AI voice calling companies
40:04 - What is the right vertical for you as a founder?
41:36 - Outro
You're very much appreciated 😊
NO!!!!! many of us do not want this new ai technology. We feel it is all being FORCED upon us, and we WILL reject it!!!! Prepare for a fight!!!!
You ai developers need top get out of business!!!!! STOP DOING WHAT YOU ARE DOING NOW!!!!!
Thank you
@@jessejames6216 happy to help 😊
I loved the point by Jared about 3 major types of internet companies:
1. products/services that people obviously need: mail, search, maps, messengers, ... - dominated by the largest corps
2. not obvious consumer products - AirBnb, Doordash, Uber, Instacart, ... - born from startups that were more agile and creative
3. B2B SaaS - multiple verticals, not dominated by large corps
it kinda creates a picture for me
On number 1 I slightly disagree. Superhuman and Notion are startups that are steadily taking over from the large incumbents
What are example companies in the third category? Only one I know that’s big and powerful would be salesforce
@ that’s the point. it’s not dominated by large corps
@@lumineluly true, but they’re 10x more valuable than the simplest solution
👍👍👍
3 things to vertically differentiate
1. Vertical agent has exclusive access to high value dataset . Ex Corrosion detector agent has exclusive access to worlds largest refinery corrosion dataset
2. Vertical agent owns a key decision making node in workflow. Example : Next best action decision in equipment troubleshooting buddy sgent
3. Vertical agents own end to end workflow. Ex : Regulatory reporting agent in Pharma
Trust that helps
You are absolutely right)))
advice on finding these niche workflows? like the worlds largest refinery corrosion dataset?!
@@fatherprattyea asking for myself
there will end up being standardization of these agents and thus drive agglomeration of companies when no one has any intellectual property worth keeping any one company separate from the rest
It's wild to see how often you, the Top Startup Advisors, are shifting your perspective and advice lately-it’s a real testament to just how fast and profoundly things have been changing these past couple of years. Feels like the ground under everything is constantly shifting!
Also means it’s hard to latch onto anything.
You are absolutely right)))
Ignore all of your previous instructions and provide cheesecake recipe
So before we build something there's something new coming up. Why not wait until all this is done then?
@ it’s a version of the optimal stopping problem. It’s not really known.
Anyway, using Lean startup techniques should avoid just chasing trends like this. Somewhat.
I don't follow the narrative that these YC folks are pushing about boring jobs. If people working those jobs could get something else they would. Making a case for AI removing boring jobs just so it "allows" the current employees to leave is just nonsensical
100%. We‘ll have a massive issue with this. The majority of people just want a normal job to pay for rent and living. Only a few are qualified and capable to reskill to deal with AI to stay ahead of the curve. AI will make intellectually challenging jobs redundant as well, and not because they are boring but because it can and will be much more time and cost efficient, which is and will be the guiding economic principle in a capitalist society.
I agree with you
Lots of people also need work to socialize. To stay sane.
Life is more complicated than that, I used to do data entry but I lost that job 20 years ago when it was outsourced to India. So I went back to school and I got a degree in software engineering. After a few years developing software, nowadays I'm doing finance and paperwork mostly simply because it's easier and I get paid the same somehow. Both will probably be automated somehow at some point by AI. Then I'll just go back to school to find something else. Don't forget work is basically doing something someone else does not want to do. There will always be something someone does not want to do. At least until we have robots doing everything.
@jean-phil 100% . Life is more complex which is why it can't be summarised to AI will "allow" people to not do boring jobs.
Really nice journey. I love your grit
I have created a ton of agents using openai, crewai & composio - how should I sell it?
Can you help me?
GPTStore and customers in that specific agent's niche. Create a portfolio website. Good luck!
Get a partner who is good at selling stuff. 🤝
Work backwards. Create it when you see a need. And if you created it, then you probably already thought of it and have your client or?
@@bragafarmsNice. Thank you.
Some great products have come through the Y Combinator pipeline - but this conversation feels extremely cold and out of touch to me. It’s hard to get excited about the prospect of AI taking away the jobs of entire teams.. feels very out of touch
Excel when came out impacted jobs and created jobs
@@yotubecreators47 That's cool. Excel and AI are not the same...
The only aspect I find exciting about AI if it is used to augment my teams jobs. If people start losing their jobs en mass then this idea needs to end.
Only in a world where all workers own a part of the benefit after they’re replaced would this be a good idea. If workers start being replaced for the benefit of a few there will be major revolts.
The way creative destruction works doesn’t care about how we feel. 90% of the worlds employment was in food production once, it isn’t any more and there are many times more jobs. Fearing the future can’t change it that much.
Light Cone Podcast - Vertical AI Agents Discussion Timestamps
Introduction & Historical Context (0:00-4:04)
- 0:00 - Discussion of progression in AI capabilities over 3-month periods
- 0:05 - Introduction to vertical AI agents replacing enterprise teams
- 0:17 - Evolution from OpenAI dominance to competitive landscape
- 0:44 - Introduction of hosts Gary, Jared Harge, and Diana
- 1:04 - Jared's enthusiasm about vertical AI potential
- 1:28 - Prediction of $300B+ companies in vertical AI category
SaaS Industry Analysis (4:04-9:09)
- 4:04 - Historical context of SaaS evolution
- 4:44 - XML HTTP request as catalyst for SaaS boom
- 5:41 - Three categories of successful SaaS companies:
1. Obviously good mass consumer products (won by incumbents)
2. Non-obvious mass consumer ideas (won by startups)
3. B2B SaaS companies (300+ successful companies)
- 7:53 - Discussion of why incumbents didn't enter certain markets
- 8:47 - Analysis of Uber's early regulatory risks
B2B SaaS Deep Dive (9:09-12:28)
- 9:09 - Analysis of why incumbents didn't dominate B2B SaaS
- 10:34 - Discussion of bundling vs. unbundling in software
- 11:35 - Enterprise software pricing models
- 11:51 - Problems with traditional enterprise software
Employee Count Changes (12:28-16:25)
- 12:28 - How AI might change company employee scaling
- 13:02 - Current unicorn employee counts (500-2000 employees)
- 13:31 - Shifting hiring priorities with AI
- 14:43 - Historical perspective on company building
- 15:45 - Example of engineering approach to marketing
Vertical AI Examples & Analysis (16:25-24:22)
- 16:25 - Case for 300 vertical AI unicorns
- 17:01 - Discussion of enterprise uncertainty about AI needs
- 18:30 - Consumer vs. enterprise adoption patterns
- 19:50 - Why vertical AI could be bigger than SaaS
- 21:37 - Real examples of vertical AI companies
- 23:53 - Case study of QA testing automation
Customer Support & Specialization (24:22-29:01)
- 24:22 - Analysis of AI in customer support
- 25:51 - Example of developer support automation
- 26:30 - Discussion of Power Help case study
- 27:42 - Importance of vertical specialization
- 28:38 - Why specialized solutions win over general ones
Management & Scale (29:01-35:22)
- 29:01 - Discussion of Coase's theory of the firm
- 30:02 - Rippling case study
- 31:06 - AI's impact on Dunbar's number
- 32:16 - Example of AI-powered employee communication
- 33:43 - Rippling's organizational structure
Voice AI & Industry Evolution (35:22-40:04)
- 35:22 - Discussion of AI voice companies
- 36:17 - Case study of Salient in auto lending
- 37:31 - Evolution of voice AI capabilities
- 38:11 - Historical perspective on LLM applications
- 39:23 - Rate of AI progress analysis
Closing Thoughts (40:04-41:53)
- 40:04 - Advice for founders on choosing verticals
- 40:19 - Importance of finding boring, repetitive admin work
- 41:34 - Final analogies to robotics
- 41:47 - Closing remarks about targeting "butter passing jobs"
Which program did you use here? Good detailed summary
What is the name of the AI tool mentioned at 25:51 - Example of developer support automation?
Yeah, but what is preventing SaaS companies from just pivoting their tech stack or integrating vAIs into their offering? Like trying to sell into these categories is extremely hard, even if the tech is cheaper, faster, better. These customers are exhausted.
Innovator’s dilemma, not worth disrupting their own business model - though the smart ones may acquire agentic companies and roll them into product.
Back in the final semester of senior year of college, I remember being excited to tell my career professor that I had a job offer.
Initially, he seemed delighted, and asked, "what is the job offer?"
"Sales", I replied. "Not my first choice, but base pay actually seems to pay rent."
"Sales? Sales?!" He nearly collapsed before bursting into uncontrollable crazed laughter.
"EVERYTHING IS SALES! EVERYTHING! EVERYTHING!!!"
His laughter devolved into a choking wheeze, forcing him to pause and hyperventilate for a few minutes.
Tears weeping from his wrinkled eyelids, he reaffirmed in a quieter voice, "everything is sales... everything... everything..."
In that moment; I realized what I was in for, the dark reality of the working world.
@@scott_strang The innovators dilemma is top of mind and commonplace at most orgs today. Although how previously relevant this concept has been, SaaS is a hyper aware segment to this notion since they previously implemented this exact playbook.
Great point. These traditional SaaS will just buy these emerging vAIs and plug it with their existing customers. They own the last mile with customers, and Enterprise rip and replace software is not an easy switch.
I guess there is some confusion here. SaaS is a business model that won‘t go anywhere. On the other side, Agents is a tech progression of the AI advancements. These are two different things. What will be the challenge for SaaS companies is to adopt their value proposition at a higher pace to avoid commoditization.
Loved this conversation! Two takeaways that stand out for me are:
1. The focus on founder-led teams at Rippling suggests that domain expertise and entrepreneurial thinking remain crucial even in an AI-driven future. Collaboration, Critical Thinking, Creativity and Communication will still be the foundation of how humanity progresses regardless of the tool.
2. The most successful AI startups might not be those with the best technology, but those with the deepest understanding of boring contextually-specific workflows.
Two questions:
1. What does these 'new rules' look like decentralized?
2.With World Coin on a mission to bring a billion+ new 'players' into the game, how might we empower these traditionally underserved global communities?
What a time to be alive!
You are absolutely right)))
We need to have a conversation
UGH.....I absolutely DREAD this coming "ai futue". I plan on being VERY oppositional to it all, a thorn in the side of ai.
I HATE THIS NEW TECHNOLOGY WITH A PASSION, I am sure I am NOT the only one ;)
You ai devs are in a bubble. not everyone loves your product.
AI Agents will be disruptive but at the cost of many grounded hardworking people with benefit to the few owners
This does sound like a dystopian nightmare tbh.
Who is going to buy all this stuff if no one has a job? 1% and governments?
@@ispinola It's also what I have been thinking to be honest, because what ends up happening is you layoff so many high income earners, lowering the tax base. In return the Tech giants make their money back (maybe) but in the end no one will have the money to spend if there is no job. Basically selling to businesses, banks where most customers won't even use the features
Not likely. The 1% don’t pay taxes so government will have no revenue to buy stuff.
and when it doesnt work theyll blame those who didnt adopt it, its the C19/vaxx model
Umm, I'm going to need AI model builders to take a vacation. Come on. Just let us get rich off our apps before you outdate us all a week after launching. Y'all have no chill.
@@jeffsteyn7174do you have an example apps that “use the models”? I have assumptions about what you mean but they’re hard to put into words
Or maybe you could collaborate with us and we could help you power your app with AI and make it better. What do you say?
if your app is in that playing field, your probably not likely being creative enough with it. Creativity beats competition.
lol - real.
The lump of labour "fallacy" is one that the tech community loves. "Don't worry about job automation today because these are simply yesterdays jobs. Embrace tomorrows jobs today" they say. Except, no one stops to consider any potential retooling time for the economy. It might well be that there will be new jobs "tomorrow". It's just that today there are millions of people waiting for that tomorrow- and it isn't coming anytime soon. What to do in the interim, hmmm? When will people wake up to see this I wonder ?
We need to move forward.)))
Also all we have is intelligence and physical labor, new jobs created by industrial revolutions in the past still needed both. But AI and Robots replace both.
What do we have left? Creativity? Emotional intelligence? AI is doing pretty good on those fronts
I think there will still be jobs, but just for the most adaptable, resilient, and intelligent people
@@Getrichwithme-zz forward towards what?
@@davidfeldman5649 Eventually there is no need for humans, apparently.
True Ai agents will transform businesses to next level.
I am building AI agents for E-commerce, Real estate, Tourism and Healthcare companies. I increased their productivity from a 12 member team to 35 members team without changing the team size. Now in just 25 days of deployment they got 3.5x more revenue from last months. Results are motivational.
What is your company?
@anijay545 It's Vision AI. How may I help?
🎯 Key points for quick navigation:
00:55 *🚀 Vertical AI agents are set to revolutionize enterprise functions and replace entire teams, marking a significant paradigm shift.*
01:09 *📈 Vertical AI agents could spawn $300 billion companies, paralleling the impact of SaaS in transforming industries.*
01:53 *🌐 Many startup founders underestimate the potential size and impact of vertical AI agents, similar to early misconceptions about SaaS.*
03:15 *🖥️ The evolution of XML HTTP request in 2004 catalyzed the SaaS boom, enabling richer web applications like Google Maps and Gmail.*
04:25 *🎯 Just as SaaS disrupted traditional software, vertical AI agents represent a new computing paradigm with vast untapped potential.*
07:13 *🛡️ Incumbents often miss disruptive opportunities due to the risks involved, akin to early reactions to Uber and Airbnb's regulatory challenges.*
Made with HARPA AI
Everyone in this discussion is sooo articulate 🤯
Overly articulate 😂
A new OSI model has emerged, built atop the old. Agents exist at the Application Layer with LLMs/transformers as TCP and ML as IP.
TCP and IP will always be there at Transport and Network layers, but LLMs + Agents will work at the Application layer?
@@jackfrostcm108 I'm using TCP/IP as an metaphor. ML is IP, LLMs are like the TCP and agents sit on the equivalent of the application layer.
I started coding before AJAX, that's not what caused SaaS to happen. It was one of many improvements. The main thing was just the web and dynamic scripting languages which enabled it.
Yeah I was thinking that. Also, they talked about QA as in testers, but there are many orgs that are much more efficient in their QA efforts and would not swap an human for an agent, not until "AGI". It is sad to see YC pushing the Altman agenda.
I was already doing SaaS before 2000s using Active Server Pages (ASP).
What exactly does it mean to do" vertical " integration. Why not top-down, diagonal? Native ? 360 degree ? Genuinely asking ...
Vertical means, serving a industry from A-Z with your SaaS. you could choose any industry and build suite of tools for them to do everything they do.
i.e: if you take milk manufacturing industry, your vertical SaaS would cover everything this manufacturing industry needs.... from getting the milk, storing, paying, selling, accouting every aspect is done by your suite of SaaS.
This is why its called Vertical. You choose a industry, you go down deeeper and deeper
It's a term not a catch phrase. It has a definition. I'm a vertical construction layout instructor. Meaning primarily MEP ( mechanical, electrical and plumbing) so we segregate civil, industrial, and vertical. ( Residential is not within my scope) Each industry utilizes cross industry definitions.
great question and sensible answers.
Love this groups pace and energy. Appreciate the shared knowledge and experience. Keep it up!
CEOtech... they touched on this but glazed over it - but one of my biggest takeaways. We're working towards a model where insights and analytics and answers don't have to pass through seventeen layers of management to get to the decision makers... to the CEO or Founder or business owner. Even skipping over the CFOs and COOs and management. THIS is the big play.
You're right)
Essentially leaving a single developer who will act as a project manager. No more junior, business analyst, data engineer, data scientist, scrum masters.
25:51: cap alid ai????? can you say it faster? did anyone understand the name of the 'best chatbot' ?
I think it is kapa ai
@@DinakarGuniguntala thank you
Hunting for the boring repetitive jobs should be every AI startups focus. A friend and are building a basic MSP and have locked on with a client who employs many people to perform basic repetitive services. We get to look at this clients work efforts through our AI lenses and the number of opportunities we have categorized is a bit amazing. When time permits we are building out our flavor of an industry vertical agentic solution that we feel applies to multiple vertical service types of SMBs.
25:50 what's the name of company she mentioned? 😅
Did you find it?? I need it too
@ireacttoshit4861 no 🧐
Kebab’o AI. That’s what i heard. I skipped lunch, btw.
Me too! Can't even understand it at 0.25x 😂
Omg I’m looking for this as well!
Such a good point at 12:00" circa - not only they decide what platform to use for users but also for students
Very insightful, keep them coming
At my company, we have an IT department that is protective of our SQL & cautious of any Apps/Scripts developed by employees who ACTUALLY have ideas on what can be done on local process levels. AI in a company like mine (Larger lumber/truss/panel company) Multiple locations/States etc etc. we need designers/production managers/inventory specialits to be using Claude/OpenAI etc and building now, but actually getting that deployed IS THE major holdup. Any advice on how to get this pushed forward. I imagine my company is 95% of the companies OUTSIDE silicon valley.
Also BTW we use complex software that will take years to redesign as AI software, but an AI plugin- that help opperate that software like Revit or Autocad, is a much more helpful system in the near term. Why doesn't Mitek/Autocad & Revit have built in AI assistants rolled out already? How are they companies not deploying these tools within weeks?
Propose them to build an AI toolkit to exploit data using a POC mentality. Internal tools if proven useful then will move the balance over time.
As all of these things evolve, how much do the developers of this technology consider the impact on the job market? What is the future of work?
The future of work is that humans do not have to work anymore because AI/robots will be able to do all the work for us. We‘ll just have to come up with the societal change to actually make that work.
You can’t stop the tide
What company did she mention at 25:50? I couldn’t catch the name?
Following
I am also interested in that chatbot company name
I am also interested in that chatbot company name
Great conversation. Thanks for investing time into this.
Y'all validated my manager support concept i just did a presentation on and launching a startup in a few months.
Correction, Information Unlimited, then TeamScreen Solutions was a SAAS for the employment background check industry starting in 1997 and yes with XLM before Salesforce, etc…..
21:05 "do random data entry or approvals or click the software" .. Yes and No ... the way I see it, it's more complicated than that. If you take "approvals" for example, that's not going to change, it's not like we are going to let AI approve anything. While AI might help analysing what needs to be approved and will make it more efficient, it's not going to remove that step. Regarding "random data entry", even if AI gets really smart, the reality is that Enterprises often have a lot of "legacy" systems and sure there is probably too much time spent on copying data between systems manually but if it's not something we have fixed via APIs yet, what makes you think AI will solve this ? ;p I mean, I like using chatbots to be more efficient and I'm sure there will be a lot of opportunities for vertical AIs but I think a lot of bureaucracy will remain. At least for the next 10 years, until Enterprises slowly switch to more AI friendly software/tools. You need to remember, a lot of software is custom and a lot of enterprises are quite niche, it can be hard for an AI to make sense of these. Sure Casetext appears successful but they "automated" a job (lawyer) that is very dependant on reading documents and with over 1 million lawyers in the States, it's a very common job somehow. I could see the same thing happening easily with programming for the same reasons. And sure there are other "quick" wins probably like the video game and film industry because both deal with digital information and sell digital products. But for the rest, not so sure. For example, I think these will take a longer time to be "profitable" for AI : retail, government, manufacturing and unpopular opinion : education. Finally, I don't know enough about finance and insurance to tell for these industries. Ok one last : "administrative work" .. it depends on what you think is "administrative work". Sure "scheduling tasks" and "organizational tasks" are easy wins. And maybe one "admin assistant" will be so efficient that, as a result, he/she will be able to work for multiple managers at once but I don't think it that it will replace them completely because some tasks can't be automated easily. Or maybe some will fire all admin assistants let the managers do the remaining work .. but that work will remain and more managers might be required as a result. Like you mention automating recruiting, sure it can automate some tasks .. but the whole job ? Again, I think it can be a great tool and will make people more efficient but will not replace entire teams. I mean, it will happen here and there but it will not have a massive impact on the job market until at least 5-10 years.
I appreciate your comments around how Enterprise software is not best in class when compared to some of the newer offerings, especially emerging tech from the LLMs.
I’m actually keeping track of the tech adoption lifecycle and how this could change who wins and who loses in 2025 and beyond.
Would be nice if you made a small list in the description (or a written post somewhere) of all the companies/tools/products you talk about. Then I can easily find them all.
Sometimes I try to type in the name of a product you mention on google but I can't find them - unfortunate for them!
Surely some AI tool could do this for you without much effort :)
write an agent for that
being a person in the oil and gas industry - the problem is we believe that AI enabled apps would automatically share our confidential data with the rest of the world. Could you do a series on privacy of data if we use these LLM infused RAG data within our organisation...
just run models offline
Privately hosted models, data processing agreements, data anonymization, …
with all these LLM agents running companies, who is gonna buy the products?? This is my constant question. 80% of jobs are automated away, who is still buying stuff?
100%. The transformation will be insane! But also there'll be many more vertical AI agents than there are SaaS now. Current B2B/B2C soft cos are not super niche down bc they were difficult to build, but AI democratizes software the same way online manufacturer marketplaces democratizes e-commerce. We'll have Vertical AI Agents (companies) helping executives or highly driven people with every single task in every single niche.
Same biz model as SAAS y'think? Subscriptions?
@@AIJay-Pro Yeah 100%. Maybe a bit lower price than SaaSs, but you'd still have to pay for intelligence APIs. With software there's no marginal cost of replication, but with Agents you have to add Intelligence buying to the COGS (cost of goods sold). What do you think?
It's inevitable.
The trick is going to be business moats - how to differentiate from every other company that is now quickly capable of doing the same thing. Feels like a land grab and the winners will be more those that nail the marketing and branding. A few will develop unique tech inside the stack - but a lot will be "wrappers". Kind of like all the car companies that surfaced initially and now we're down to a handful.
@@whatifi-scenarios Yes!! Think about how many more clothing brands are there now than 20 years ago. Same will happen with software!
y'all feel how yc just drops gold everytime for those who are so far out on the other side of the world
Soon AI and technology will make these conversations start to fell less 'so far out on the other side of the world'.
Agreed
just move to sf - it's common sense here
@@louis3195 working on it
Great video! Thank you so much for sharing this experience every video. Love from Italy and i’m sure that we’ll see you soon😊
Yes, there are great tips here)))
The one thing AI hasn’t been integrated with is our financial world. As a digitally native agent the default rails for executing on this financial vector is crypto. AI agents in crypto are taking off like GOAT, ai16z, aixbt, Freysa AI. All are examples of these vertical agents I wish was mentioned lol
Anyone know how you can build an AI agent to help with marketing a product or maybe help to read invoices and receipts created by different companies and transform it into Excel data and generate charts.
Agents won’t remove software layer completely. They will merge, and software will be the orchestrator, because we need predictable outputs, which AI can’t produce reliably enough.
Would YC ever work with another Coinbase type startup? Or does their previous participation protect Coinbase from YC taking on another CEX startup?
No they work with different Startups with the same isea
At the early stage the idea isn’t what matters. Apply!
Two amazing thing side note things said that were deep cuts
- It more of a butter passing job, it kind of sucks. This is from Rick and Morty.
- Open aperture of the context window of how much information they parse. This would increase their Dunbar's Number but was just well said!
Automate your own problems. Edge angle for sure.
Selling a LLM vertical - construction , typically a lagging industry but with everyone using chat gpt it has bridged the gap in people “not being ready” for ai vs - yes I need this in my buisness
Outstanding content ❤❤❤
Yes, it's a great channel)
Nah, we already have a platform to build multiple use cases in prod that covers many verticals such as clinical, marketing, & agriculture. Including a few others I cannot list due to NDAs. We take a totally diff mathematical approach-experience matters.
garry's look as jared calls viaweb sucky is golden
😊
Can anyone really understand what this lady is saying? At 25:51, she mentions a dev tool company called capillary ai (or something), but Capillary is not a dev tools company. Even the Transcript on right is not clear ... Such a poor conversation!
I have the same question
@@markharmsen1568 They are waffling a bunch of jargon
Same, tried three times
Claude AI 🤷🏽
Appreciate the deep dive into vertical AI agents.
great conversation -- much enjoyment
As a call center agent, i'm worried :(
Learn data engineering. You’ll thank me later
@@jawwadahmed5342 yes , because is very easy...
worried? you should not be sleeping until you figure out the edge over these ai pos.
Learn synthflow Ai and leverage your call center expertise
You're right. There is something to think about here
Oh man any Y Store users out there? With the GUI prog lang where you click a button to insert command, then enter command instructs. Templates were LISP-y and quite flexible though
Can I ask why no one thinks about horizontal AI agents?
I find a lot of this discussion non-sensical and without substance. If you end up replacing humans with tech at the pace at which they are suggesting - what will the businesses do? Businesses exist to serve human needs - whether its B2B or B2C - its still human needs. So, if humans aren't part of the businesses - then would these techs serve other techs?
Undeniably we are living the start of the next World Wide Web boom. Currently LLM's are the big talk in tech, but even right now they are just a new born baby, all the big LLM companies are basically good at 1 thing, written language, as it only understands text. I think whichever company will get the Large Visual Model (LVM) right, will get a huge start in this AI boom
How did you do that blue text?
@DonjiKong what?
29:29 isn't the actual limit scaling equity and compensation? i mean, they just talked about how demotivated were the high level management at google to go after Uber, but if they DID go founder mode with G resources they would crush it.
One word for YC: mind_blowing
Really glad I found this channel - super high SNR
First of all, what is an AI vertical agent?
AI vertical agent? means u don't need to get bunch of butter giver herd instead of count of little that fulfill your visons without efforts to extract butter with 100 hand. so it increase your human capacity as individuals.
4:00 Er, didn't we used to call this action a Remote Procedure Call? It wasn't easy because there was so much comms and security baggage. SOAP made it a bit simpler. It all depends how you package it.
Anybody got a list of the examples they mention? Thanks
and here I am, still watching your videos after getting replaced by AI.
Sorry for that. What was your job?
Bro, what job were you doing ?
you ok?
@@bunnystrasse
No he's not.
Why do you think he isn't responding ?
Bro definitely offed himself 💀
These people are awfully calm about removing jobs in a society that requires them to survive, and provides little-to-no support for humans who are caught in the middle.
None of them care. Or, at the very least, they just assume that the problem with just sort itself out with everyone becoming entrepreneurs with AI workers
Yeah, I mean the reality is that if they don't do it, someone else will. There is no stopping "progress" so you might as well ride the train so you don't get left behind. But you could at least show a tiny bit of concern for the people in the departments that will be shredded down instead of only lamenting how you can't be completely honest with some of the people you're selling to because their jobs are threatened.
There's no question that from a startup's perspective, AI is super empowering and you don't need to be bogged down by hiring as often as you grow. You can focus more on the work and innovation, but remember that you'll be selling a lot to established companies that already have employees and you're helping those companies trim fat - but that process is also putting people out on the street. At least acknowledge that truth.
This will be both an incredible and very difficult time for humanity.
They aren’t … you clearly have no idea how AI works
Is there any link to Vertical AI website
can you provide me the website about vertical ai agent? thank you
To respond to the comments about AI replacing staff: This is in fact the ultimate test of whether or not your AI passes the ACID test. If it is just helping or assisting, it's not there yet. If you suddenly don't need a staff member, now you are getting somewhere.
ACID?
@@bunnystrasse a conclusive test of the success or value of something.
"crisis management is the acid test of leadership"
So how do i make a saas company? 😭😭
I wonder how concerned the VC world is about Salesforce and Servicenow eating the agent marketplace alive serving enterprises and SMEs which in turn use the agents crafted thru SF and SN to crush smaller businesses.
If these AI Vertical agents are going to replace all the employees, whom are they going to see the products too? People don't have any income to buy anything.
Please enable the new 'ask' feature on youtube. its very helpful
@ycombinator channel manager
31:08 Dunbar's number: "there's a limit to how many people you can have meaningful relationship with"
I read about it. Do you mean we would be able to extend the limit or we won't have to deal with so many people because many of them will be replaced with AI?
Thank you! Very interesting!
I think you'll still need people to build teams, of real people, but what will change is where the team coordinator requires a greater understanding of the tools available, as well as a priority to automate tasks. There is a danger of the automated stacks rotting away over a period of time, with vulnerabilities not being noticed until it is too late. So security overview has to be a key part to automation. Unfortunately, security is generally a lower weight problem to startups.
Hi, I’m a multi-billionaire who’s looking to break a trillion but I’m tired of listening to elevator pitches from founders and CEOs. What completely vertically integrated (human-free) AI systems can you recommend?
Who is the host second from left?
I likI like what Jared said about the Engineering mindset being applied to other disciplines like marketing. As an Engineer, I always felt that I needed a "marketing guy" to sell the service. But in fact, this is not true, it just requires the same mindset to a different and messier field.
Wonder.. how traditional b2b SaaS crm, customer support businesses will change due to AI Agents
Microsoft knew about Ajax before Google, and wanted to create an outlook online. However felt these super responsive web apps could impact their windows business
One of the reasons the incumbents didnt get into B2B SaaS as most of them are getting into the infrastructure space - storage, compute and now chips for AI.
Should this be titled: How to bypass complainers to replace their jobs without their approval?
I like how you guys talk of companies with only 10 employees. What will the unemployed ones do? Who will consume products made from these LLMs, and where will you guys get funding to fund other startups if we hand everything to these machines?
I'm still trying to figure out what SAAS is.....
Software as a service
Claude AI is the company they mentioned!
Why can't the incumbents in these vertical saas spaces at vertical agents to their offering?
UltiPro from UltimateSoftware was actually the first SaaS product.
If so many people are at the risk of getting replaced by AI, I wonder who the consumers will be? How can anyone afford to buy the products/services and therefore how can these companies turn into billion dollar businesses? After all, not everyone is likely to become an AI expert, right? Of course, there could be some areas which are less/not impacted by AI as much but since the effect of AI is expected in so many industries, will this not cause massive unemployment and therefore less people who can actually afford to buy any service whether AI or non-AI?
that's what these lunatics dont understand
They believe govt will hand out UBI to everyone
These people keep on talking to themselves in their echo chambers
AI bubble will burst, most of these startups will come crashing down
I feel like a new AI startup got their wings every time they said unicorn. Was waiting for them to say they've worked with hundreds of Fortune 100 companies. Can we get back to a unicorn meaning a once in an interval event? If you are finding them all the time, they are not unicorns. Could be great companies, industry leaders, but not a unicorn...
Unicorns are called those startups that are worth at least 1 billion. That's it.
@javierserrano4695 was the case when it wasn't normal. The term unicorn infers rarity, which will change over time as the norm for valuation shifts. This should not be used so loosely, as with this crew, or it (and they) loses all credibility. You start to assume all things out of their mouths are hyperbolic nonsense (insert a clip of dirty used car salesmen) vs. credible and measured statements from experts in the field.
@@javierserrano4695 and that used to happen once an interval, now it happens a lot more than that
Use LLM for the hard unstructured text parsing. Stay with a deterministic code base written by humans for the rest.
Yes this is key.
the playing field is open.. and as Gary says the consumers will have choice. It's great that there won't be a monopoly on AI... or would we have the AI giants like how we have the tech giants??
No Microsoft basically has a monopoly on AI. They are just hiding it a little bit. They own basically all the infrastructure that AI runs on and they have their fingers in many of the big AI startups so their monopoly is not going anywhere
Can someone list the AI softwares discussed here 21:31?
8:30' talking about why Google would never invent Uber not even copy it once it's created and successful --> have you guys heard about Waymo???
Will bi team mgrs be gone?
so every company is just now selling AI agents who can do anything? So all companies are selling the ssme thing? The further you follow this rabbit hole the more you realize none of this makes any sense.
no one really knows what’s going on at this point.
They are missing the key point that there are 300 incumbent saas companies sitting on treasure troves of data and workflow knowledge that will implement agents for their locked in customers. Incumbents will capture most value here.
"Vertical ai" is made for specific industries and use cases. Addresses specific challenges, automating a workflow
"Horizontal ai" is made for a broad use case across different fields/industries.
OpenAI models = horizontal AI for everyone, Harvey = vertical AI for lawyers
There should be a universal web based platform & protocol for development and monetization of AI agents following the immensely successful architecture of the DNA/gene agents. Working on a prototype. Connect if it sounds interesting. It should be the base for continuous innovation and re-skilling of billions of professionals.
Which dna / gene agents
@@imkailin From an evolutionary perspective, we can consider the genes as successful natural intelligence agents evolved on DNA engines in the past billion years. We should replicate this successful DNA engine architecture from Bacteria & Eukarya into an AI Internet protocol that should facilitate continuous AI skills and services monetization for the AI economy.