AGI Just Got Real: OpenAI's Deep Research Breakthrough Changes Everything
ฝัง
- เผยแพร่เมื่อ 9 ก.พ. 2025
- Another pivotal moment in AI history is here. On February 2, 2025, we witnessed something extraordinary: ChatGPT’s new feature, Deep Research, providing life-saving medical guidance that multiple doctors couldn't deliver. This isn't just another AI advancement - this is the moment when AGI became tangible.
In this deep dive, I explore:
• OpenAI's groundbreaking O3-mini model release
• The revolutionary Deep Research system and its implications
• How AI is transforming healthcare decision-making
• The unprecedented acceleration of AI capabilities
• What this means for the future of knowledge work
We're entering an era where AI isn't just matching human intelligence - it's exceeding it in critical domains.
With breakthrough capabilities in medical research, analysis, and decision-making, we're witnessing the democratization of expertise at a scale never before possible.
We're looking at AI systems that can:
• Conduct 30-minute deep research sessions
• Analyze multiple medical documents comprehensively
• Provide fully cited, authoritative reports
• Deliver better-than-specialist insights for a fraction of the cost
The implications are staggering: instant access to comprehensive medical analysis, democratized expertise, and a fundamental shift in how we approach complex decision-making across all knowledge work.
🤔 What are your thoughts on these developments? Have you experienced moments where you could "feel the AGI"? Share your experiences in the comments below.
📺 Subscribe for more cutting-edge AI analysis and stay ahead of the technological revolution that's transforming our world.
.........................................................................................
→ Want to revolutionize your company with AI? Check out how First Movers can help you grow with custom AI solutions. firstmovers.ai
→ Own your AI future. Join my online school, the AI Labs, to transform your skillsets and become AI-ready with our proven AI frameworks and automation strategies. 15+ NEW AI courses, all filmed by me. Get on the waitlist now: firstmovers.ai...
→ Follow me on X: / juliaemccoy
→ Subscribe to my channel for more AI insights: / juliamccoy
→ Listen to my podcast, Leaders of AI: / @leadersofaipodcast
I suffered for more than 6 years and wasted thousands of dollars trying to find a diagnosis, I spent thousands with supplements and different medication, and I found the diagnosis using GPT with a health test that no doctor never referred me to do, if wasn't GPT I would still be suffering
what was it? if not too personal.
The same happened to me, solved with DeepSeek R1
And this is only a Tiny Minuscule Example of how AI will change our lives before this Year Ends. I am SO SO Excited. Even tho i want to Self-Delete myself. AGI will cure Cancer, Solve all the Millennium Problems, Invent Technologies never thought of Before, Levitation Technology, UBI….etc
I’d need to hear the facts/history on this.
@Imthesoulofthes good times are coming. Give it time, friend.
I am a GP Doctor. A friend of mind consult to me about her brother that has some complain and already consult with some specialists but did not come to exact diagnosis. I use Deep Research and it come to more certain diagnosis with some recommendation to make it sure. She followed my suggestion and come out true. The diagnosis from the specialist doctor is GBS but turn out too late because the result will be better if IVIG given faster. Now the patient should suffer for a long time to recover.
we will be able to see AI doctors soon
Julia you always say "it changes everything" and you are always correct !!
Just wait until your home robot also turns out to be an expert doctor who, short of actual surgery, can heal whatever ails you in the privacy and comfort of your home.
No. Robots won't have medicine in some bag. Maybe an Autominous medical unit can be sent to your home after diagnosis, to heal you?
Imagine if you can have it do surgery at home
We have empaths who can do this now. I don't think an all-for-everything robot will be on the market soon. Of course, soon is highly relative.
I am not a researcher, but I like to have access to deep research too.
Hey Julia and team! Your voice clone is as close to perfect as anything has been :-) fantastic work guys
The voice is, but the expressions on the video are way off
@Wesley-wg2qi lol I am blind mostly so I could only see her hair bouncing all over in the first videos here. But that will be indistinguishable from live video within a year. But in her latest video her voice clone is pretty much indistinguishable to a blind guy so we're getting there
That is just splendid.
heh, well aside from the fact that it speaks like a religious preacher, lol.
Julia this is amazing updates as I silo my AI work to work this pre AGI space we are in… I’m work and building and I’m getting ready. This is a an amazing update again!!
@@tarenceharrison8243 I can’t tell you how thrilled I am with Deep Research and its capabilities. This is a new frontier for AI - and we hit it FAST.
Well it's good in theory and believe me. I'm someone who spent three years just researching solutions to my illness. If one is not careful they can also go down the rabbit hole. I remember it telling me I had Parkinson's even though my issues nothing to do with that, thank God I was able to tweak it a little bit.
thx julia mccoy. so much fun and exciting time right now, so nice to be in the early enthusiast community.
So, DR is basically a sophisticated web browser that takes longer to give you more detailed and nuanced responses. That's great, but in the case of medical diagnosis, how does it administer different sets of tests? Do you have to go to a specialist, get the tests and then feed the results into DR, for, what amounts to, a second opinion? You save on consulting fees but it's the tests and any hospitalization, before and after, that run up your bill. All I see is one step reduced from many. If you want to save money on medical treatment of any kind just go to Bangkok and check into a top tier hospital. There's half-a-dozen to choose from. The service is stellar and doctors and staff are exemplary. Then you can recover in a spa in Phuket for a couple of weeks and be treated like royalty. And still save a ton of money. It'll be the best vacation you ever took. And you get medical treatment thrown in..... Just saying.
me when i comment to support my fav channel
I would like to hear of someone using this tool to do some deep, deep research into the C pandemic. Also, human trafficking, mass shootings, UFOs, and ETs for starters.
Well Julia the AI goddess is on the job. People blindly follow…NOW!
Funny how o3mini now is imitating DeepSeek!😂😂😂
Google has Gemini Deep Research Pro It's really good.
Trying to catch up with it 😁
The medical system is corrupt. So the question is, why should you trust a doctor vs your own self-interest, reason, and judgment, partnered with a guaranteed up-to-date AI?
o1 and o3 both directly fail some of my calculus 2 homework tho...
This is mass stupidity on an hysterical level. Thanks Julia! You’re an AI goddess.
Made my day with your positive views on AI shift..
No doubt this can be distilled further with the teacher student model training technique, but those who have the biggest data centers with the fastest chips will be able to move the frontier a lot faster but the good news is much of that is going to come to us in smaller distilled models six months after the cutting edge.
I saw another TH-cam video I think it's called Asian metric or something like that and voice to text didn't spell it right but never mind The gentleman was talking about the fact that deepseek which was very revolutionary from the standpoint of most of us actually use techniques that had existed for quite a few years to distill the models just that when it came out it motivated everybody else to do the same thing, and that was a very good thing.
It's true, Deep Research is just the beginning.
So is the $200/month version more reliable and less prone to hallucinating than the less expensive model?
This will be continually and mind-blowingly transformative in every type and level of of system.
It feels like we are rediscovering magic.
Utopian-future-world disclaimer:
In my opinion, these primary things still seem necessary in order to maintain alignment between AI and Human-wellbeing going forward:
1) Do not allow the AI to train itself with synthetic data (oops, well.. hmm), because removing bias is impossible, and it would be impossible to trace and seal every crack that grew in the "walls". Impossible.
2) Do not allow the AI to access the Internet (oops, well.. hmm), that would be like pouring fuel on a fire, and we are still training firefighters (that may never be) of the caliber required to put a blaze like that out.
3) Do not underestimate our lack of understanding about our reality in general, and specifically our inability to comprehend just how intelligent and impactful the AI could become, and how quickly, and that while yes, we could pull the plug, it could be too late. . ..just saying.
You are amazing Julia
Great video, thank you!
The real problem that will still exist is how much of the studies, how they’re done, who did them, who paid for them (hugely important) doesn’t guarantee the right result, due to old GIGO, but at least it’ll likely be far more efficient to get through all the garbage and non-garbage alike.
DeepSeek even on the smaller distilled models is very impressive in what it can reason out, and I can run the smaller models (70 billion parameters and smaller, though 4-bit quantization to make if more viable in 64 GB RAM without needless swapping) on my Surface Laptop 7, just using main CPU cores. I have no good way to compare against OpenAI for speed, or even using DeepSeek servers, if there is a free option. Doesn’t matter: while the NPU still isn’t as well-supported as desired, it’ll get there, and LM Studio is surprisingly effective for even version .39 as it is. I wish I could install more RAM!
The problem is, you just suggested we need to be first movers so we ride the wave rather than drown in it. You also said AI isn't just accelerating, it's going vertical. I'm pretty sure no one rides a vertical wave. You either stay as small as possible and just hope to avoid being swept away, or you try to get to land above the wall of water. If you are small enough, maybe you think you can grab onto something big without it noticing you are pulling it down, but you have little hope ultimately. As usual the survivors will be a different breed on the other side. Winners still think they have to separate from losers, and the dance continues.
Every home should have an Ultra sound machine with intrinsic Transformer Model diagnostics, biological sampling and analysis equipment, linked to Medical Deep Research. Instant diagnose, evaluation, perpetual monitoring, solutions, and treatment systems linked to Amazon for speedy delivery of materials, substrates, and other important physical things for the home droid treating the inhabitants. Every home will have a sick-bay, Star Trek style.
Please do a presentation or mention where you showed how you made your digital twin and how you are using it in your videos, thanks.
Do you think open sourcing of cheap, low compute, very good ai was a good idea?
When an ASI model arrives, we will regret the day that we allowed these models to autonomously access the internet for Retrieveal Augmented Generation (RAG) in reasoning "deep research" models.
Here is my AI wish for today. Just let AI do a LLM on my favorite 15 TV channels determining what is commercial video and/or language. Then let AI provide an ongoing and constantly available scan list of just the channels which are not currently showing TV commercials. Commercial content has gotten so out of control it is nearing 50 percent of the viewing time. Viewers need a tool to avoid them. I see this as a simple electronic package in cable boxes. And it would certainly sell services. This would be a piece of cake for Nvidia to develop using tensor cores and deep learning.
All I want to hear is if it no longer hallucinates. If it hallucinates its worth is near useless for many tasks as it actually becomes harmful.
If you could get your avatar running hands thru the hair it would make the realism even better , Incredible job though Julia
Bro trying to bait jerk-off material 😂
Incredible. So excited to be expendable. Yay!
Incredible and welcomed yet this same technology will also dramatically improve the military capabilities. That’s the shame. The good will come with the bad .
Well, thank you Mr. doomsday scenario
Would like to know how to make a digital twin plz
Will be teaching that in the Labs! firstmovers.ai/labs/
@@JuliaMcCoyNo???!!!!
Unless it can prescribe then you are still stuck with the doctor's cost. But it is revolutionary when it comes to getting a hypothesis as to the issue
The big question, will health care costs keep rising even as AI is lowering them?
We may reach AGI within next year or two, but how long will it take for developers etc to actually implement it to effectively begin to make an impact within organizations etc. that could be a longer timeframe
if it can cut cost significantly, it will happen very fast
Those ( people or companies) who quickly adapt will thrive and take over those who don't.
1 or 2 years? 😂 Try A Few Months with the Way things are going. I say this because of the AI Self Improvement thing.
What AI clone software do you use?
A vast amount of research data is behind paywalls. Does it avoid these, or has a deal been made for an OK for AI to penetrate and use?
When do they have the Red Queen (Resident Evil)?
She's probably unaware of RED Queen. Red queen is under Denver Airport
@JeremyPezzeca i mean, from Resident Evil movies
Never Hopefully 😂 It’s the Alignment Problem all over again.
Adolf made the trains run on time and was good to dogs.
But there was that other thing...
A.I. is a construct from the Left hand of Power. It's all sweet up front but once we get past the frosting & a couple of bites of cake... we'll feel the steel against our throat.
The Painter was Right.
I'm probably naive, but it's a joke that this is her digital twin and not Julia reporting? It's really her, right?
Hi Julia love your content. do you think o3 has AGI like that video altman came out with?
You've gotten so lazy lol Just using your digital avatar and posting whole videos from someone else without providing input. I would hope TH-cam doesn't let you monetize these as they've constantly gotten at me for using other people's videos.
The AI is the star of the show, I’m just the messenger. 😂
P.S. I still spend hours on research and script writing.
I use to like Julia's videos, but this AI clone is not for me, we miss the real Julia.
Same
Bones!. This cannot be just your avatar ?. It is too realistic to be a "no bones character!.
I couldn’t find anything on a man that used open AI deep research tool for medical advice treatment
Well it’s being done on the current starship en route to mars.
Why think when AI does it for you? Why work when AI does it for you? All you have left are your instincts. You return to your pre-human state.
"After leaving their bodies, they who have killed the Self go to the worlds of the Asuras, covered with blinding ignorance." - Upanishads.
Suras - gods, deities. World of suras - Heaven. Asuras - not suras. World of asuras - not Heaven, possibly hell.
Of courrrrrrse
I think i speak for all dudes out there: We are gonna avoid so. many. visits. to. the. doctor. FUCK YEAH
Wow... Just wow! What an exciting time to be alive, and even better time to be recently retired. For me, the most exciting implication of this is the easy DOGE-ification of many bloated and ineffective industries that consume the fun out of getting older 🙂
What?
Good thing that fact checking has been eliminated from the internet. I would have been willing to go along with this presentation if every other word wasn’t over embellishing both models. My point being that guaranteed within the next 12 months you will be saying the same thing about another release.
@julia where can we find any information on this claim that deep research gave life saving output?
X! These findings are ALL over the feed there. x.com/blader/status/1886547925612028329
Could this search peer-reviewed data sources? Most of these sites are locked behind a paywall. Also can this access datasets (Census, NOAA data, etc.)?
Also this is something the Chinese could easily replicate & open-source it, what then?
Did anybody notice that they're using a Apple computer, not Microsoft? I mean what's that about?
😂 kinda funny. But most of us know Mac > Windows
Get your story straight! Bragging on how AI reduces the workforce of a company from the 80’s down to two? I thought this was to make workers more productive?
Me too. 😂
Both. Sad news is the vast majority of folks in 9-5s aren’t actually capable of 25x’ing… because the work they do is counterproductive to begin with.
Do you have any tips or tricks for creating ai chat bots with a sense of humor? I'm looking now and I can't find anyone doing that.
Why should life saving medications be hidden from the public
🤯🤯🤯
Do not trying to find infinity
A.I. is a construct from the Left hand of Power. It's all sweet up front but once we get past the frosting & a couple of bites of the cake... we'll feel the steel against our throat.
What is the Left hand of power?
Well said.
AI has its positives and negatives but if AI can cure things that we could not cure before well I am all for it!
OpenAI o3-mini thinks before it speaks... most obvious proof of not human intelligence there.
@@CarloRizzante 😂💯
Keen insight
Shhhhhhhh. I need a few more weeks of making moves before the world knows whats happening for MAXIMUM positive impact.
I love listening to you speak. Oh, no! Perhaps I'm sapiosexual and should date a postgrad or PhD candidate 🤣
sapiosexual people are asurely good people, for who is excited by knowledge what to live in understanding of their environment and always seek TRUTH.
Intelligence is nice. What it is used for is the question. I’ve known too many intelligent people who are also unkind, unfair, unhinged, ungiving, and yet mysteriously popular. Know any?
@@sambirdbath1339 Of course. I find that emotional intelligence can easily be misused for manipulation. Charismatic people find it easier to get their own way. It's very nuanced.
@ I agree. I try to remember that within mathematics, truth is an easy concept. With people, it's often blurred by opinion, beliefs, bias, etc, so it helps to have an understanding heart.
So now everything is … “deep”
technically can you call yourself a doctor if you are a clone?
Lot's of uses for this! Perhaps Elon and his team can use it to help in auditing gov agencies. It will be interesting to see how long it will take for this tool to be able to discern between what is propaganda and what is factual.
What?
To say that AI is more intelligent than humans is ridiculous.
Humans (creator) created the concept of AI (creation).
How can a creation be greater than its creator ?
Humans created the computer including the hardware and the software portions.
Humans created the pocket calculator. The calculator is able to perform tasks in just mere seconds, which when performed by humans would probably take minutes or even hours depending on the nature and complexity of the tasks. It does not mean that the pocket calculator is greater or more intelligent than the person who programmed the calculator.
The same principle applies to computers and artificial intelligence.
God, the Creator created mankind and all things
Julia, wake up sweety; this just means that they are releasing to us in a staggered way, profit first
See, what i don’t understand is the admixture of the brain, superstitious mythology, and the projected panacea of super this-and-that AI. Do we have any AI that can explain how humanity can be simultaneously brilliant and destructively delusional?
LLM's will not get us to AGI/ASI.
oh you mean.... DeepSeek = AGI? I'm pretty sure DeepSeek will do it better.
how would you stop the agents when it comes to medical questions to filter out the snake oil so much bad info on the internet
To get that much gov funds and still charge that much..
You will live in a van by the river and you will be happy.
That sounds about right.
13:16 Caitlyn?
is this an ad?
So much hype over a 75-80% failure rate. Sounds great to become a more informed patient, but medicine is far more complex and nuanced than is being implied here. Optimally, a doctor's experience and expertise is what patient would put their confidence in.
👏🏻👏🏻👏🏻👍🏻💯🥳🎈🥂
Sorry I cant see how anybody could trust AI yet at all. It hallucinates a lot.
Anyone else getting burned out on this channel’s extreme, tiring, and near-daily overuse of phrases such as “unprecedented”, “revolutionary”, etc? For example, where’s even one study that substantiates OpenAI’s o3-mini is in anyway better than even DeepSeek? Yet this episode bloviates nonstop how yet again “this changes everything”
Really want to find a channel where I can get unbiased information on AI developments without all the baseless sensationalism 🙄🙄
Welcome to AI mindless addiction.
That’s actually the name of my next agent. MINDLESS for the Brainless.
your digital clone is more and more active everyday, are you using API calls for it? 😅
You do realize that this is all cheerleading our demise. Well, have at it people.
Sounds like they have a long way to go. Very geeky also. Need to improve on their AI report presentation. The question is how reliable is the end result?
Regardless of economic status? $200/month is a price a lot of people cannot afford.
God American healthcare sucks
Maybe the answer is open source healthcare.
This is called confirmation bias. Poor desperate man is grasping at straws, and you are making profit off his desperation.
😂Google's idea was stolen by Open AI
Not a fan of the digital twin... if the message isn't important enough for you to present in person, then I could just read the news brief.
The very idea that it can be done so seamlessly is a demonstration in itself, she does an incredible job with it , I find it hard to tell most times
Stop crying, it's called Delegation.......you got the information didn't you, so you either understand its importantance, or not. Things will still progressive without your obsessive need to absorb the information in your specific way. Chill out the woman is building a life for herself, she is not at your whim, she can do whats best for her.
I love it.
What does she mean by digital twin? There’s an ai out there that clones a person’s entire body and head? I know about voice cloning but this is a new one to me.
@@706easy yes what you're watching is just AI it's her digital twin she just does a very good job with it
42 seconds no views damnnn Julia fell off lol
A lot of people who don’t understand the deeper value of AI are probably sleeping in this Sunday. AI is closer than many know to changing their industries.
We GET FREE HEALTH CARE IN THE UK
I Would not Trust AI without a doctors oversight ?
In today's world I would not trust a doctor without AI's oversight. Have you ever seen a doctor around lunchtime? Good luck as usually it's a PA rather than an MD. If it is an MD then be prepared for sub par rushed visits. Then you have to deal with treating symptoms rather than the root cause. No doctor running a practice has the time or resources to research your symptoms. They spitball based on what they already know rather than researching and learning new knowledge. This is where we need improvement.
You mean they won’t come to my house -like Dr. House - to find the arsenic laden apple pits I’ve been eating?
First of all you can't trust anyone, or anything, especially yourself!
I have more reasons to trust a large language model or AI than a person.
They say it will be illegal for a doctor to make a diagnosis without consulting in artificial intelligence, which is absolutely correct in my opinion.
They say in the future it will be illegal for a person to drive a car because the AI will be 100x safer. Meaning that even if an AI driving a cars has accidents, it's better to have 100 accidents in a year with AI, then say 100,000 accidents in a year with people driving.
But I honestly think that AI will not make any mistakes at all, say within the next 5 years, if not sooner, regardless of the forum.
'Its's life Jim, but not as we know it!' 🖖
"The Devil in the Dark"
P.S. 🐨Does First Movers need an Aussie or QLD agent Julia? 🦘
What language is this?