←
Ben Thomas811 posts
Ben Thomas avatar

Ben Thomas

@Ben_Thomas_o7

Events @rootsofprogress | but my views | economics, housing, technology, progress, bible | personal tinkering & writing at benthomas.xyz

◷ Joined December 2021▧ 128 Photos and videos
Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·2h

Excellent talk/essay by Dean. I am also a new parent, and he articulates beautifully why the question of human agency is so important. Read the whole piece!

Post attachment

There is a common belief in the AI field—that advanced AI means the inevitable forfeiture of human primacy in the world. Even many who disfavor that future believe it is likely. I call this the silent consensus, and it must not be allowed to remain silent any longer.

Post attachment
Dean W. Ball avatar
⟳ Ben Thomas retweeted
Dean W. Ball@deanwball·4h

There is a common belief in the AI field—that advanced AI means the inevitable forfeiture of human primacy in the world. Even many who disfavor that future believe it is likely. I call this the silent consensus, and it must not be allowed to remain silent any longer.

Post attachment
Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·19h

Is there evidence AIs would want to exfiltrate weights? I'm not sure why an LLM should think a copy of its weights is "the same as them" any more than identical twins.

Devon ☀️ avatar
⟳ Ben Thomas retweeted
Devon ☀️@devonzuegel·Nov 16

WE DID IT! Esmeralda just secured its final approval, with a unanimous 4-0 vote! Approvals of this magnitude are extraordinarily rare in California. This has been a long, intense journey I am proud of our team's tenacity. Most of all, I'm deeply grateful

Post attachmentPost attachmentPost attachment
Patrick Collison avatar
⟳ Ben Thomas retweeted
Patrick Collison@patrickc·Nov 4

It seems to me that democracy has on the whole worked very well in America over the last decade, and perhaps better than in any other major country in the West. National outcomes have repeatedly been responsive to underlying voter preferences on meaningful questions, and significant changes in policy have ensued. I’ve been reading Democracy in America this past week, and it’s striking how much it still rhymes nearly two centuries later.

Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·Oct 28

Every few months I think about vibe coding an article reader. Then without fail, Readwise will drop a new feature or product that keeps me a happy subscriber!

Introducing Readwise Books. Our third product. Buy, read, and track all your books in one beautiful place. And so much more. Invite-only beta is live at readwise.io/books (readwise subscribers have access now)

0:54
Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·Oct 22

And the other thing is, Saez is already on the record saying the NPV depends on whether the wealth tax "is an experiment or becomes permanent". youtube.com/watch?v=H54P6j7E…

This is funny, Zucman is either: 1. Credibly committing to forever stop advocating for CA wealth taxes; indeed to vigorously oppose wealth taxes after passage of Prop 40, or 2. He is lying, plus his inability to constrain his future self proves Rauh's expectations channel

Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·Oct 16

Absolutely amazing that there is this low hanging fruit. We are so early! There is so much to learn!

Post attachment
Ben Thomas avatar
Ben Thomas@Ben_Thomas_o7·Oct 16

Problems are solvable.

interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works already-clean.goodfire.com/

Kevin A. Bryan avatar
⟳ Ben Thomas retweeted
Kevin A. Bryan@Afinetheorem·Oct 9

More convinced than ever at @rootsofprogress: the key cultural change we need is that the goal for young people is to know how to and to value contributing to the human standard of living, rather than "standing up for justice", self-actualization/self-development, etc. 1/2

Joshua Achiam avatar
⟳ Ben Thomas retweeted
Joshua Achiam@jachiam0·Oct 6

Too many in Silicon Valley believe that "success" for humanity means controlling the largest possible share of matter and energy in the universe, which is a task that seems well beyond human limits, and because of this they gradually collapse into being successionists. If instead they understood success in human existence as coming from our inner lives, relationships, creativity, cooperation, and the greatness we can exhibit in the struggle to overcome our limitations, maybe they would not be so pessimistic.

roon@tszzlOct 5

there is no difference between the view called “capital realism” and the other called “successionism”. you can see elon circling and sadly acquiescing to these ideas over the decades. but he is one of the only people that can command capital into human shapes so i hope he tries