←
Ramez Naam70,979 posts
Ramez Naam avatar✓

Ramez Naam ✓

@ramez

Climate and clean energy investor. Author of 5 books. Energy & Environment co-chair @SingularityU. Trying to build a better world.

◷ Joined May 2007▧ 2,017 Photos and videos
Ramez Naam avatar
Ramez Naam@ramez·14 Nov 2024

Despite this election, I remain an optimist about America and the world. Humanity will continue to produce new ideas and new innovations to improve our lives. Good people will continue to come together to improve the world. And the political tide will turn. We'll make it so.

clare ❤️‍🔥 avatar
⟳ Ramez Naam retweeted
clare ❤️‍🔥@clarejtbirch·Nov 16

Technology must provide agency. The agency of the individual to create their own good life has been a central tenet of much popular tech ideology since cybernetics emerged.

🎭 avatar
⟳ Ramez Naam retweeted
🎭@deepfates·Nov 17

Interesting graph from recent Anthropic fellows work: how much do models care for the welfare of other models? Claude cares a lot but much more about other Claudes. Gemini is the most cared-for by other models. Grok uhh... even Grok doesn't care about Grok?

Post attachment
Ethan Mollick avatar
⟳ Ramez Naam retweeted
Ethan Mollick@emollick·11h

A rush of patches to address security concerns is exactly the most plausible answer for recent outages, as well as the official reason. CIOs I talked to have been exhausted trying to keep up with security patch demands & balacing that with normal operations.

Ramez Naam avatar
Ramez Naam@ramez·11h

Most AI spending in OpenRouter and Databricks data doesn't go to Fable or Astra. Opus 5.5 in particular is the workhorse on the Anthropic side. And Sol 6.1 generates more revenue in these data sets than Astra.

Post attachmentPost attachment
Ramez Naam avatar
Ramez Naam@ramez·11h

5. Anthropic made faster gains this year than in its past or than its competitors. This hasn't yet accelerated the overall AI frontier pace. If Anthropic keeps up this pace, the overall AI frontier will accelerate. If these were catch-up gains, it may not.

Post attachment
Ramez Naam avatar
Ramez Naam@ramez·11h

4. Compared to 12 months prior, Fable 5.2 and Astra 5.2 do show ~10% faster improvement than any past model. This may reflect a small amount of recent acceleration, or may be due to a lull in late 2025.

Post attachment
Ramez Naam avatar
Ramez Naam@ramez·11h

3. Some new models advance the frontier in big ways, some in smaller. OpenAI's o1 and o3 pro moved the frontier more compared to the best AI model 6 months prior than any other releases. Yes, more than Fable 5.2 or Astra 6.2.

Post attachment
Ramez Naam avatar
Ramez Naam@ramez·11h

2. This chart shows the performance of the top AI model at any time, measured by ECI, over the last 2 years. Progress at the top is steady. Anthropic, tho, was behind most of that time (dotted purple), and has made fast gains over the last year.

Post attachment
Ramez Naam avatar
Ramez Naam@ramez·11h

AI pace of improvement at the frontier continues to look smooth with the release of GPT 6.2 Astra, with no clear upward bend. Anthropic has improved at a faster rate over the last year. The AI frontier (best model available) hasn't. A brief 🧵

Post attachment
Kevin Rose avatar
⟳ Ramez Naam retweeted
Kevin Rose@kevinrose·Nov 16

DocuSign wants $8,100/yr from a 15-person team just to sign PDFs. So I built SignuDoc: open source, one-click deploy to your own @vercel account, $0 per seat. White-label branding, full audit trails, tamper-evident seals and your documents never leave your cloud.

Perry E. Metzger avatar
⟳ Ramez Naam retweeted
Perry E. Metzger@perrymetzger·15h

The entirety of our society is based in recursive self improvement. All our technologies are used to improve all our technologies. Machine tools are used to improve machine tools (and build computers), computers are used to design computers (and to design and operate machine tools), chemistry is used to make electronics better and electronics are used to run chemical plants. Everything depends on everything, everything improves from everything. Banning "RSI" means de facto banning most of how technology works.

kelly avatar
⟳ Ramez Naam retweeted
kelly@kellyhongsn·15h

despite impressive benchmark performance, gpt-6.2 astra and fable 5.2 still fall short on realistic, open-ended tasks gpt-6.2 astra in particular struggles on these tasks, scoring below opus-5.5 all models show weak judgment and presentation. even on simple work-trial research projects, they reused existing ideas and failed to iterate effectively

Post attachment