Gavin Harris (@GavinFHarris) profile picture

Writing on SI at the Foundation for American Innovation Views my own

Washington, D.C.
Joined October 2026
Photo by Gavin Harris (@GavinFHarris)Photo by Gavin Harris (@GavinFHarris)
"the most important AI R&D contributions aren't yet fully automated" = claude can't yet do long-term strategy, judgment, and risky innovation consistently? Phew, was starting to worry that McKinsey consultants wouldn't have any reason to exercise their creativity...
Replying to @METR_Evals
3. Interviews and surveys with staff are consistent with AI usage driving current acceleration, though this evidence is very hard to interpret. Staff self-report dramatic productivity gains, but say the most important AI R&D contributions aren’t yet fully automated.
1
54
placing my chips on Estonia as the new Singapore. In a few years Estonia could be the first AI-governed state 🇪🇪
"We are already a digital nation, and now we want to become an AI nation. We have the mindset, the people and the digital infrastructure to make it happen." — @KeitKasemets, Secretary of State 🇪🇪 🔹 Fireside chat: OECD Secretary-General and State Secretary of Estonia #TDS2026
2
33
Just dropped AI → SI → AI → SI in the chrome web store chromewebstore.google.com/de…
💀 someone created an SI → AI chrome extension to counter our AI → SI extension @GavinFHarris you know what to do next
2
1
9
3,097
One insane bit from today's amended Anthropic S-1 I anticipated below - $100.2 billion in (minimum) compute commitments signed since September 30! That's $100B in a month. Also, $895B confirmed total compute commitments (up from $807B that I totaled) policygradients.thefai.org/p…
2
6
823
To make the transition from saying AI to SI as easy as possible, we are releasing an “AI → SI” chrome extension that automatically swaps "AI" for "SI" and "artificial intelligence" for "super intelligence" wherever you go on the internet! Get it here: chromewebstore.google.com/de…
50
8
24
71
52,557
You've probably already seen something on Anthropic's S-1 on your feed. It's definitely weird, but I decided to zoom in on the two things you need to know to understand Anthropic's prospectus: the governance structure and the compute commitments.
1
2
2
8
2,936
Surprising tidbit #3 After 10 years, successors lose their seats and the founders are unable to choose successors. Founder control also sunsets automatically if fewer than 3 founders remain after 90 days.
1
2
33
What to watch before the IPO: the vacant board seat, the blank compute totals, changes to the Founder LLC or Trust, and how Anthropic handles cancelable contracts like SpaceXAI. Full breakdown on Policy Gradients, co-written by Govind Pimpale: policygradients.thefai.org/p…
2
33