Nothing Much Happens in AI, Then Everything Does All At Once
When it rains, it pours. OpenAI Operator tested and reviewed, with full paper analysis. Perplexity Assistant is useful. Then Stargate, is it all smoke and mirrors? Strong rumours of an o3+ model from Anthropic. Then a full breakdown of Deepseek R1, and what it’s training method says about the state of AI. It’s not open source BTW. Plus Humanity’s Last Exam, and Hassabis Accelerates his AGI timeline.
AI Insiders ($9!):
Chapters:
00:00 – Introduction
00:54 – OpenAI Operator
04:53 – Perplexity Assistant
05:15 – Stargate
07:51 – Better than o3?
08:25 – DeepSeek R1 Analysis
12:12 – Training Secrets
15:19 – No More Process Rewarding ?
19:01 – Hassabis Timeline Accelerates
21:22 – Humanity’s Last Exam
System Prompt:
OpenAI Operator:
System Card:
There is No Plan:
Perplexity Assistant:
Stargate:
Labour goes to 0:
Noiam Brown Manhattan Project:
Larry Ellison AI Surveillance:
Amodei 1984:
Microsoft Hesitate:
Dylan Patel o3+ for Anthropic:
Deepseek R1:
Diagram:
Process:
Demis Interview:
Humanity’s Last Exam:
Non-hype Newsletter:
Podcast:
Hello ,thank you for all the videos.
Hi just wanted to say I’m always wanting for your videos and entering YouTube for that most of the time
One of these days we’re going to need to be enhanced by AI to keep up with all the AI news, lol. Happy we have *you* now though, haha!
Hi. Would you have a moment to tell me how you make some text boldface in a comment? Thank you. 👋
Woohoo! 23 mins of quality content! 😁
Thank you for this video. 👋 Like some of you, I’ve been thinking about and researching advanced machine learning, longevity science, and robotics for over 45 years. (I’m 66) I no longer can afford a serious life extension program for myself, and despite all the developments I see happening in this tumultuous era, I expect to die before they can extend my life. Somehow, I am able to sustain a melancholy, but positive mood enough to be grateful for levelheaded AI news analysis such as provided by this channel. I’m grateful for it. Anyway, it’s just me and my dog. I’ll say goodbye now. Thanks for the efforts.
Even though I’m always already aware of what’s going on, it feels great to get the infos in a rich video format. Please keep the videos coming, I’m always looking forward to your uploads, Philip! 🙂 Also, since they already have o4 and are working on a Level 6 Engineer, and also considering the first big Texas data center is almost finished, they’ll train GPT-5 on valuable synthetic data generated by o-models increasing its capabilities while also profiting from ~50x to a ~100x compute EOY 25, making the coming frontier insanely valuable in regards to new releases like operator-like agents. You can feel the pull – we’ve already crossed the event horizon, it feels like.
“You don’t invest in that many charging stations for EV’s unless it’s going to be the next big thing.” said us 4 yrs ago
Great video, keep these coming!
Best AI channel 🙌
Bruh I worked in IT HUMANS aren’t smart enough to get out of failed loops
This is brilliant as always: not a minute wasted. Favourite nugget of insight: “why should chains of thought be in English, *or any human language*”?
I will have to say, the definition for AGI that Demis used in that clip is effectively ASI under most understandings of that term. So, what he is saying is that we will have ASI in 3-5 years. This matches with Dario’s prediction in his recent interview, predicting for by 2027
Why reward every correct step when you could simply punish the incorrect ones? It can be easier to determine when a step is wrong, rather than when it’s right, and using punishment would allow multiple approaches to be taken. Punish wrong steps, reward the right outcome.
Again, your videos are appreciated greatly. But I wanted to ask about something: in the first section in which you go over OPERATOR, you suggest that once an open source model is available and becomes greatly improved because of its chain of thought aspect, you say that the Internet will be changed forever. I’m sure you’ve explained that in prior videos, but I’m frustratingly wanting to know exactly what you mean by that. How will it change forever? Obviously I have some intuition about what you mean, but it would be great to hear the specific dystopia you’re likely picturing. 😮
Love these videos, best ai channel on yt
can’t wait for the deepseek version of this 1:19
The day I can get an AI to work with me using the entire context of my github project, is the day I will celebrate.
If my model doesn’t know extremely obscure details about hummingbird anatomy, I don’t want it!
Deepseek also closed source their custom AI training and inference framework which lead to such efficiency gains. If they didn’t omit anything, they fully explained their method. But in the meantime they have an edge (specially noticeble on deepseek’s inference pricing vs companies like deepinfra or together ai).
Sure I like these videos for the content but also for Philips’ friendly and enthusiastic tone of voice and the beautiful British accent