
Climate and clean energy investor. Author of 5 books. Energy & Environment co-chair @SingularityU. Trying to build a better world.
Seattle
Joined May 2007
- Tweets71K
- Following10.2K
- Followers62.8K
- Likes211K
Pinned Tweet
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.
Ramez Naam retweeted
China makes 97% of the world's synthetic graphite for battery anodes.
As of today, a plant in Oakland makes it from natural gas, with 3x less electricity and zero process emissions.
Ramez Naam retweeted
9/ One prediction hasn’t shown up in the data just yet: that AI displaces entry-level jobs first, because juniors overlap most with what the models already know.
So we still have time to invest in “flight simulators for work” that get the next generation to the frontier.
Humans continue to be the bottleneck on AI improvement.
Replying to @ccatalini
3/ The tension isn’t any different at the frontier: at Anthropic, where Claude already writes more than 80% of the code, the company says human review has become the key bottleneck.
Verification is increasingly the bottleneck on AI productivity and impacts on the real world. Great thread.
Replying to @ccatalini
2/ AI coding agents have drastically accelerated code creation. But what reaches production hasn’t moved nearly as much, and about half of what passes automated tests can't be merged.
Security is worse: in CMU’s tests, about half of the top agents' working code is insecure.
I love this. Way to catalyze this, @CJHandmer
Today I’m launching what I think could be an extremely important research project.
Fixing humanity’s scientific knowledge.
The idea is inspired by physicist Casey Handmer, who used AI to screen more than 50,000 papers in theoretical astrophysics. His audit flagged 872 possible errors.
This is impressive. But why stop at finding the errors? Let’s use AI to FIX them.
And why stop at astrophysics?
The ambition of this project is to go through every scientific paper ever published, across every scientific discipline.
AI agents will review each article and produce a revised version that corrects the errors they find.
The aim is not to shame authors or seek retractions. We are humans and we all make mistakes.
The ambition of this project is to correct all the errors that we have made so that current and future generations of scientists can build on firmer foundations.
I’m looking for researchers from every discipline to help make this happen. If you’d like to take part, please fill out this short survey and tell me which disciplines you’d like to help cover.
Feel free to share with anyone interested.
Survey link: unimib.qualtrics.com/jfe/for…
The risk of over-indexing on fear of AI is that we give up freedom in exchange for perceived security. This manifests in Paul Christiano's recent interview. It also manifests in _some_ proposed safety policies that would surveil and restrict the ability to compute.
Ramez Naam retweeted
AI has caught up with climate change on people's lists of extinction threats:
outpaced.substack.com/p/ai-h…
Some quick reactions to the recent Dwarkesh podcast with Paul Christiano. Tl,dr: bullish on rapid capability gains, more skeptical of the leap to takeover, and wary of emergency overreach.
Ramez Naam retweeted
Interesting to look at it this way, silicon commitments by the frontier labs. Cembalest: "Nvidia has only a 39% share of future frontier lab compute agreements compared to its 80%+ market share from 2023-2025... here’s evidence that some competitors may be closing the performance gap on inference workloads"
Ramez Naam retweeted
this is like, the best cyber news we've gotten in a long time? things working
NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say
AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems
washingtonsun.com/technology…
Ramez Naam retweeted
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.
Ramez Naam retweeted
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?
Ramez Naam retweeted
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.
NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say
AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems
washingtonsun.com/technology…
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.
OpenRouter: www.teapawt.org/PeterJ_Walker/status/2…
OpenAI vs Anthropic wallet share just before these most recent GPT-6.2 launches.
Anthropic: 64% of spending between the two labs, led by Opus 5.5
OpenAI: 36% of spend, leading model 6.2 Sol (6-Astra as a close second)
1
4
994
Databricks: www.teapawt.org/pwendell/status/212237…
1/ Today we’re publishing the Databricks AI Index, showing trends from enterprises who manage AI with our Unity Gateway. Data covers >10,000 companies with broad representation (Digital Natives, Fortune 500, Global 2000, etc...). Enterprise AI consumption is changing quite a bit:
1
743
Ramez Naam retweeted
A good list! The 10 accounts that think most like me:
@krishnanrohit @entirelyuseles @AlanMCole @herbiebradley @sebkrier @Dan_Jeffries1 @ramez @robinhanson @martin_casado @gbrl_dick
via @neighbors__ai
meetyourneighbors.ai/u/binar…
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 🧵
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.
1
15
2,734





































