Building @gen_intuition: models for envs that require deep spatiotemporal reasoning. I like games, OSS, AI, and once built the world’s largest RuneScape server.

New York, NY
Had a great chat with one of my favorite investors, @generalcatalyst General Intuition x General Catalyst - The Next Phase of Pre-training
“The digital world lacks physical information. The physical world lacks dense digital information. Games perfectly merge these two together, and we believe that’s just the next phase of pretraining.” Every few weeks @gen_intuition ships new emergent capabilities that are orthogonal to anything you see in the LLM world. Watch GC’s @max_rimpel in conversation with @PimDeWitte. Chapters 00:00 — Introduction 00:10 — The World's Biggest Private RuneScape Server 03:09 — From RuneScape to Ebola 07:41 — Mapping the Unmappable 09:47 — Why LLMs Can't See the World 13:45 — The Accidental Foundation of General Intuition 19:10 — Turning Down a Life-Changing Acquisition Offer 21:12 — One Foot in Front of the Other 24:04 — Atoms to Atoms 27:33 — The Talent Flywheel 30:20 — Protecting the Last Weird Corner of the Internet
8
5
53
61,280
Obligatory @tbpn round announcement with some spicy questions on the AI consciousness and humanity debate from @johncoogan (No I do not think they are😅)
2
1
35
8,652
Pim de Witte retweeted
Happy Friday. On today's show: - @jgebbia (U.S. Chief Design Officer) - @gpcastle12 (Anorak Ventures) - @PimDeWitte (General Intuition) - @themollette (Dorm Room Fund) - @cityofthetown (Vanity Fair) See you on the stream.
BYOCompute, Pope Destroys False Idol, Ben Affleck's White Pill x.lingyaoai.com/i/broadcasts/1qxoNYjAj…
3
6
36
23,855
Pim de Witte retweeted
It’s quite disturbing that Anthropic has been running a campaign to lobby the major faiths to adopt their own philosophical interpretation of AI, rather than taking input from them. I’ve been fortunate to spend time at the Vatican, and I’ve always asked for advice and guidance, answered the questions asked of me, and been grateful to them for their time and guidance. It’s insane to me Anthropic would actively impose their viewpoint to the extent of threatening to pull out of events that don’t precisely align. It’s a moral arrogance and dogmatism that has characterised EA and Anthropic (certainly its leadership at least) for a very long time. I’m glad these stories are coming out and being exposed.
Note well, apparently Anthropic tried very hard to get the Vatican to alter the text of "Magnifica Humanitas" just prior to its release so that it would not deny AI personhood.
26
55
561
31,128
This is a good explanation of what we do at @gen_intuition, and why it’ll be a big unlock. We train general agents who are great at pixels to actions, partially using world models as a way to scale data and capabilities. If you’re interested in joining us, reach out!
10
11
113
9,291
My version of buying a nice car 🥹
6
43
2,485
Turns out scaling on data and compute with a great team is surprisingly effective 😉
General Intuition has raised another $220M at a $6.2B valuation as we start making our models available. These are an entirely new class of foundation models, trained on billions of action-labeled videos, capable of acting in realtime in never seen before environments. 1/6
24
9
203
16,851
One way of viewing simulation theory is: we’ve been given a sacrificial universe without asking for anything in return, which can calculate at speeds beyond our comprehension, and we are to make the most of it. Most religions are built on less sacrifice. Yet we call that love.
1
1
14
1,984
Pim de Witte retweeted
Through the VOICE trial, Terry is using his Neuralink implant to help fine-tune a brain-to-voice interface for himself and others who can’t speak. He trained the algorithm first by miming speech as best he could, then by simply thinking the words and hearing them come out in his own natural voice. Powered by Grok Voice from @SpaceXAI
697
3,234
20,064
3,443,130
Pim de Witte retweeted
Last month I wrote about how we can build a positive and safe future for everyone: meta.com/thefutureisforevery… Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens. The reality is: - People won't want to use agents that are misaligned with them and that don't do what they ask, so labs have a strong natural incentive to make their models more aligned. There is a lot of debate about slowing progress on capabilities until alignment catches up. My view is that trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models. Any lab that doesn't focus on alignment will fall behind. - Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well. Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us. I'm proud of the security foundations we've built. - Engaging independent evaluators and advisors is industry best practice. MSL already does this today in several areas because it helps produce better work. Other labs can just do this too. In general, it would be helpful for there to be a larger and more diverse ecosystem of evaluators. - Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely. Meta has made this commitment and other labs can do this as well. I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do.
1,928
2,861
28,451
7,750,967
Has someone tried putting the fly in a world model yet
5
1
45
35,552
Love this
How can we reconstruct a memory that was never filmed? Our team paired restored archival photos with pose control models to capture the mannerisms and micro-expressions of Burt and Ethelle. This helped bring the day they first met to life for Love, Rendered, a new documentary in collaboration with @PrimordialSoup_ and @StorySyndicate_. Watch the full film on YouTube → goo.gle/3V8m94m
32
8,456
Pim de Witte retweeted
Worth the read every year. #neverforget Todd: Hello… Operator…listen to me…I can’t speak very loud. – This is an emergency. I’m a passenger on a United flight to San Francisco.. We have a situation here….Our plane has been hijacked…..can you understand me? Lisa: (exhaling a deep breath to herself) I understand… Can the hijackers see you talking on the phone? Todd: No Lisa: Can you tell me how many hijackers are on the plane? Todd: There are three that we know of. Lisa: Can you see any weapons? What kind of weapons do they have? Todd: Yes…. they don’t have guns….they have knives – they took over the plane with knives. Lisa: Do you mean…like steak knives? Todd: No, these are razor knives…like box cutters. Lisa: Can you tell what country these people are from? Todd: No…..I don’t know. They sound like they’re from the mid-east. Lisa: Have they said what they want? Todd: Someone announced from the cockpit that there was a bomb on board. He said he was the captain and to stay in our seats and stay quiet. He said that they were meeting these men’s demands and returning to the airport… It was very broken English, and… I’m telling you…it sounded fake! Lisa: Ok sir, please give me your name. Todd: My name is Todd Beamer. Lisa: Ok Todd….my name is Lisa…Do you know your flight number? If you can’t remember, it’s on your ticket. Todd: It’s United Flight 93. Lisa: Now Todd, can you try to tell me exactly what happened? Todd: Two of the hijackers were sitting in first class near the cockpit. A third one was sitting near the back of the coach section. The two up. front got into the cockpit somehow; there was shouting. The third hijacker said he had a bomb. It looks like a bomb. He’s got it tied to his waist with a red belt of some kind. Lisa: So is the door to the cockpit open? Todd: No, the hijackers shut it behind them. Lisa: Has anyone been injured? Todd: Yes, ..they…they killed one passenger sitting in first class. There’s been lots of shouting. We don’t know if the pilots are dead or alive. A flight attendant told me that the pilot and copilot had been forced from the cockpit and may have been wounded. Lisa: Where is the 3rd hijacker now Todd? Todd: He’s near the back of the plane. They forced most of the passengers into first class. There are fourteen of us here in the back. Five are flight attendants. He hasn’t noticed that I slipped into this pantry to get the phone. The guy with the bomb ordered us to sit on the floor in the rear of the plane……….oh Jesus.. Help! Lisa: Todd….are you ok? Tell me what’s happening! Todd: Hello…..We’re going down….I think we’re going to crash……Wait – wait a minute. No, we’re leveling off….we’re ok. I think we may be turning around…..That’s it – we changed directions. Do you hear me….we’re flying east again. Lisa: Ok Todd…. What’s going on with the other passengers? Todd: Everyone is… really scared. A few passengers with cell phones have made calls to relatives. A guy, Jeremy, was talking to his wife just before the hijacking started. She told him that hijackers had crashed two planes into the World Trade Center……Lisa is that true?? Lisa: Todd…..I have to tell you the truth…..it’s very bad. The World Trade Center is gone. Both of the towers have been destroyed. Todd: Oh God —help us! Lisa: A third plane was taken over by terrorists. It crashed into the Pentagon in Washington DC. Our country is under attack….and I’m afraid that your plane may be part of their plan. Todd: Oh dear God. Dear God…….Lisa, will you do something for me? Lisa: I’ll try….if I can….Yes. Todd: I want you to call my wife and my kids for me and tell them what’s happened. Promise me you’ll call.. Lisa: I promise – I’ll call.
Community note
No verbatim transcript of Todd Beamer's call with Lisa Jefferson exists, as the conversation was not recorded and the text includes fabricated dialogue. snopes.com/fact-check/tra… usatoday.com/story/news/fac… nps.gov/flni/learn/his… catalog.archives.gov/id/2610189
101
1,329
10,257
1,748,825
Potentially the first tweet in history to cost $1T or more (despite probably good intentions) Explain working here to your family anywhere except in SF now. If anything goes wrong this will be referenced back during the trials. Good luck
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
6
1
74
14,464
When I was at MSF during Ebola, and during more recent disease out, I often found myself wondering how scientists, who are (still) often silenced in their own countries for their views, can more openly share information. I thought of launching this at the time, but didn’t. 1/
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
6
1
27
6,197
6/ once it gains traction institutions can get their own accounts and facilitate official responses to keep the discourse even and fair. There are probably a lot of lessons to carry over from public health. Thinking more about it every day. Sharing to discuss.
1
2
443
7/ this will also allow for effective company to company channels (private boards I.e. between top labs and/or governments to specific groups of scientists) and eventually the goal would be for it all to be a public discourse end to end.
324