fan of OpenAI and Ilya Sutzkever, not affiliated with anyone from OpenAI; will one day master the AI, starting from bottom but will reach the top some day
This has the most alpha per minute of anything you’ll see on mainstream TV maybe ever. Karp came on and spoke straight truth, facts and deep relevant perspective that the main stream audience has no insight into and it’s awesome. There’s really nothing to agree or disagree with on here, just learn, that’s how great it is.
Sam Altman in the financial times:
“In another year or two, we expect to have built systems with astonishing power, capable of delivering tremendous value to the world. Artificial intelligence will reshape the material conditions of human life on a scale that no technology has accomplished since the harnessing of electricity, and perhaps beyond even that.”
I obviously would conquer and believe AGI as I define it - capable of replacing the majority of humans in white collar roles - to arrive in 2029.
Remember OpenAI is targeting a GPT 6 in August, which will beat fabled 5 in all benchmarks. Then a few months after that we will see another step change. This year will be much more exciting than 25’
Sorting which financial docs are worth an analyst's time is surprisingly hard for frontier LLMs. With an expert-labeled dataset and on-policy distillation, Bridgewater fine-tuned a model to do it reliably and cheaply.
https://t.co/gyYzXq15zd
What does the next training paradigm look like?
0:00:00 – The big research bet the labs are making
0:02:12 – Grindability is just as important as verifiability
0:06:10 – Will RLVR alone generalize?
0:08:41 – Getting the learning back to the weights
0:15:22 – Dreaming
0:17:23 – What 2027 looks like
Also on YouTube, pod feed, and Substack.
Beff, from how many of my tweets you have liked, I know how badly you want to be pals. But I am sorry, I can never miss an opportunity to remind you that I think you're a huge jerk and that your reactionary and destructive e/acc-holery has been a real factor in driving polarization against tech (though admittedly a small one because it hasn't breached containment into the real world yet). You had every chance to be a responsible and welcoming ambassador for technology, but you chose instead to antagonize people trying to be good, and you chose to alienate the broader public. I am perpetually baffled by the choices you have made; you have hurt the cause of progress and you do not seem to have learned any lessons. I hope your company succeeds because the tech is cool and good tech should be successful, but I really hope you become less of a wad.
“Mathematicians and scientists often speak very different languages. AI is potentially going to lead to a real renaissance in applied math, where pure mathematicians who are domain experts are now going to have the perfect conversation partner to be able to take their ideas and connect them with real-world things.”
https://t.co/MecehlhW9a
How close is AGI?
@OpenAI Chief Research Officer @markchen90 discusses what the future of model capabilities looks like: "We're getting closer and closer to a world where the models can come up with more of the innovation on their own."
Really fun to hang again with my friend 🃏 @polynoamial (OpenAI research scientist, our first guest ever on @NoPriorsPod in early 2023) to talk about the implications of large test-time compute, and what happens when models are given $10M budgets to spend on a single task. Topics:
01:23 – Why Benchmarks Are Broken
04:19 – Compute Budgets and Projections
06:48 – How Long Should Models Think?
08:01 – Benchmarkmaxxing
09:48 – Noam's Evals
12:40 – Safety (When Model Capability Scales With Spend)
16:09 – Implications For the Model Release Cycle
18:34 – Latent Model Capability
22:27 – Limits on Recursive Self-Improvement
28:38 – Large-Scale Multi-Agent Coordination
30:39 – Competition at the Frontier
33:19 – Breaking the Benchmark Grid Equilibrium
34:57 – Why Benchmarks Should be Scaled by Cost
new: Anthropic’s critics say the company is becoming dangerously powerful.
I spoke with former staffers about how technological dominance, a $1 trillion business, and political influence fit into the company's larger goal: guiding the world safely through transformative AI
OpenAI research scientist, Noam Brown:
"there's this hypothesis of an overnight intelligence explosion where models become superhuman in an instant, and i don't think we're headed to that world"
The biggest bottleneck for all of us is time, and time limits what we can do
Codex usage at OpenAI gives us a preview of what agentic work may look like in the future.
In a new paper, the OpenAI Economic Research team looks at the broader shift from chat to delegation: people using agents not just to get answers, but to hand off longer, more complex work.
https://t.co/CQzNwpEevY
A super long overdue (3+ years?) post on scaling laws.
Compute is expensive. Scaling laws are a way to help us reason about the optimal compute allocation between data and model size before committing to a large run.
The post covers what scaling laws predict, how compute-optimal allocation works, why Kaplan et al. and Chinchilla disagree, and how data limits + fitting details make extrapolation tricky.
https://t.co/HP26eJvjHB
13/ Our analysis suggests that AI demand is more revenue-validated than any prior platform shift. The investment case comes down to whether falling prices can move enough token volume to earn a return on CapEx.
FULL REPORT: https://t.co/BoYCeZDZYm
Thanks to @alexolegimas @jaimesevillamol @shanumatthew93 for early feedback on the report.
Work at OpenAI is being transformed by agents, in every department.
Across our entire company, people are using Codex to do work that is more complex, longer-running, and increasingly cross-functional.
Our internal usage offers an early look at how agentic tools may reshape work as they become more capable and broadly available.
Tim Cook, who told The Wall Street Journal that the jump in costs was unlike anything he had seen “in any area in over 40 years.”
Biggest price jump in anything I’ve ever seen too. https://t.co/aypJGgssnN
We had a fascinating conversation with @tylercowen about how he uses ChatGPT to follow his curiosity wherever it takes him. He calls himself an “infovore” with ambitions to become an “information trillionaire.” Read here: https://t.co/sQk3fqtCmO