AI – 2024AD: 212-page Report (from this morning) Fully Read w/ Highlights
Just 14 highlights from the 212-page report (in its 7th edition) on the State of AI in 2024. I add my own analysis, bringing in the latest reports of OpenAI behind-the-scenes finances, Chai-1, a new AI Consciousness job listing, the Nobel prize winners speaking out, my own predictions, and much more. Lead author of report: Nathan Benaich.
AI Insiders:
Chapters:
00:00 – Introduction
00:43 – o1 changed the game
01:37 – OpenAI costs revealed
02:39 – MovieGen and Pika 1.5
03:39 – Nobel Prize Winners + job listing
06:51 – BrainLM
08:20 – Open Chai-1
08:48 – Stacked Accelerations
10:10 – China Enters Stage Left
10:55 – 100x Drop in Price? Or 80%? + Transformers
12:16 – Copyright Models + Zuck Warning
14:12 – Interesting Climate in AI
16:48 – Jailbreaking still an Issue
17:42 – GenAI Actual Harms DeepMind
18:37 – Predictions
State of AI Report:
Pike 1.5:
OpenAI Projections:
Hinton Interview:
Alphafold 3:
Chai-1:
BrainLM:
Calliope Networks:
NYT Authors:
Zuckerberg Warning:
Schmidt on Climate:
Altman and Sutskever 2023:
AI Cooling: .
Your Bills will Rise:
GenAI Harms:
Hassabis Warning:
Consciousness Google Job:
Jailbreaking Amodei:
My 2023 State of AI Video:
My Coursera Course – The 8 Most Controversial Terms in AI:
Non-hype Newsletter:
I use Descript to edit my videos (no pauses or filler words!):
Many people expense AI Insiders for work. Feel free to use the Template in the 'About Section' of my Patreon.
Excellent insights, as usual 🙂
Thanks Mark
Streets ahead indeed 👌
😂👍
I guess the reason behind Hopefield and Hinton winning physics is simply because cutting edge physics is at a stand-still atm. There is nothing Nobel-worthy the past couple of decades in pure physics that hasn’t already won. I think it was a good pick. AI is used heavily for analysis in practical physics.
AI will enable physics to continue to advance.
@@squamish4244 but Nobel prizes aren’t awarded for future physics progress
Frontier Physics research lost touch with reality about 30 years ago.
Einstein won the 1921 Nobel Prize, not for his theory of General Relativity but for his explanation of the Photoelectric effect where he tied theoretical early quantum mechanics to practical experimental observations.
Today cutting edge physics has become a branch of pure mathematics, looking at String Theory, M-theory and a whole host of other TOEs like loop quantum gravity none of them have practical experimentally testable proposals and won’t for centuries if ever.
The problem with ‘AI outsiders’ winning the Physics Nobel is that it’s not for actual Physics research as linked to direct experimental observations, it’s for basically the development of algorithms to process and interpret data.
I know some physicists who are completely gobsmacked at this choice of winner, not because these awardees were not physicists but because the award is for something completely unrelated to basic physics as an experimental science.
Thanks for being the best source of easy AI news
💯 agree
Im reading “easy” as “reliable“, which is also “refreshing”.
Cheers both
💯 agree x2
Ever since AlphaGo and other projects completely defied expert predictions by multiple decades, I’ve realized even the experts have no clue what the future holds with this technology. Truly, we are at the base of a massive exponential trend that will likely peak mid to late century, but at this point it would not surprise be if that peak arrived much earlier than anticipated.
Yeah the problem is, if you’re an expert you know based on current rates adjusting for specific KPIs, you have a good estimate. This doesn’t factor for groundbreaking leaps in technology though, this is why AGI has gone from even in 2019 55% of experts thought AGI will occur either after 2060, or will never occur.
Kurzweil always knows better than engineers
“Aspirational brainstorming notes” 🤭😊
Excellent work, Philip. Cheers!
I thought you dont like AI no more??!??!
@@dad1844 he never said that, simply not gonna make videos on it
@@Akuma.73 except when he does, of course.
Here come the personal chief AI officer. Pfffft
Hope you’re doing well Dave. Cheers
It’s very clear to me that Ilya was the heart and soul of OpenAI while Sam was PT Barnum..
It’s so much worse than that.
@@spoonfuloffructose PT Barnum was a horrible person, so…
Ilya didn’t want to give AI to the masses tho
about the break-out part on the 10th prediction. A quick google search gave me this definition: “General Usage: In everyday language, a break-out can describe a sudden and notable event or achievement, such as a breakout performance in a movie or a breakout hit song.”
“… absent an invasion of Taiwan by China.” … that this is not hyperbole, but a necessary qualification for your prediction says a lot about the state of the world.
wait why? taiwan and china have beef?????
@@sometard8156 bro does not know
@@sometard8156 tsmc is probably the only thing stopping china or inversely it maybe even the reason in the future
@@sometard8156 Taiwan make computer chips. Future of AI requires them. Whoever controls Taiwan controls the future of AI. Whoever controls the future of AI controls the world.
@@SirQuantization First you get the sugar, then you get the power, then you get the women.
Thanks so much for your work here. The jailbreaking aspect is one I hadn’t been keeping myself refreshed on in particular.
omg I love all the Pikaffects throughout 🤣🤣
The guy is talking about AI being smarter than us in 5-20 years? Mate it’s been smarter than me for two years now lol
I still don’t get why he’s being so conservative on when that will happen. He does come out and say 5 to 20 years but 5 is less likely. I’d say 5 is a good guess but it looks bad to be wrong so maybe he doesn’t want to commit.
Because he’s not overhyping it I guess? Sure it can do any exact science better than you, but it gets hard time understanding what happens when you turn a cup with a ball upside down (yes I now, o1 can sometimes guess this correctly now, you get the point).
When they say smarter they mean absolutely smarter in any way possible.
Bro, it doesn’t know how many R’s are in the word strawberry. How stupid are you?
At the moment AI is like a calculator. It can do many tasks way way better than humans. For example solving PhD level physics tasks. If AI were smarter than humans though, why would OpenAI and Google not replace their workforce with AI?
@@drakey6617 something to look out for in the coming years
Running an auto-encoder on brain stimuli is kinda crazy
My friend you are crazy fast with this stuff
phillip ai continues to be the frontier model for parsing and analyzing the latest papers in AI news
“Streets Ahead”, a fellow community fan 😂. Great work as always.
20:30. “Break-out” status, would mean it goes main stream, much like when Chat-GPT broke out of the academic circles and into mainstream consciousness.
AI Explained: the ONLY channel where I click like first, and then watch all the way through, every time. Great work, Philip.
Thanks Rob!
YouTube played an ad just this channels said that it was about to play a video showing the AI generating audio at the same time as video… it actually took me a second to realize what had happened (but not before I thought: “damn… this is really good!”)