Paul Meinshausen (@PMeins) profile picture

CEO/Founder at Aampe, angel investor (@supabase, @Bifrost, @locofy_ai, et al), earlier CoFounder/CDO @ PaySense (acqd 2019). Compute x Chance.

Singapore & San Francisco
Joined February 2012
Photo by Paul Meinshausen (@PMeins)Photo by Paul Meinshausen (@PMeins)Photo by Paul Meinshausen (@PMeins)Photo by Paul Meinshausen (@PMeins)Photo by Paul Meinshausen (@PMeins)
As Hume summed up Berkeley's (ironic coincidence of name 😅) arguments: they "admit of no answer and produce no conviction"
I think there is about a one in three chance that within the next year, AI companies will build superintelligence: that is, AI that is much smarter than humans in every way. If this happens, I think these AIs will probably take over: that is, they seize control of human institutions like governments, or they destroy those institutions. I don’t want AI takeover to happen for several reasons, but one of them is that there’s a serious chance that as part of it all or most humans will be killed (either immediately or within a few years). I’ll say a bit more about each of those claims. It’s a bit complicated to explain why I think that superintelligent AI within a year is ~33% likely. The main source of evidence that affected me recently is research released this week by METR, an AI research nonprofit, which finds that current AIs are now good enough to increase the rate of AI progress at Anthropic by about 70%. It’s very hard to increase the rate of AI progress by this amount. This implies that the AIs are making AI research faster by a factor much larger than 1.7x; maybe it’s as if the researchers are doing their work ~10x faster. You probably couldn’t achieve this speedup even if you hired every qualified human in the world. When you plug this number into mathematical models of AI development, this implies that superintelligence next year is plausible (and that superintelligence within the next 3 years is more likely than not). See the modeling by the AI Futures Project for more on this: blog.aifutures.org/p/q35-tim… I think that superintelligent AI is likely to lead to AI takeover because AI developers do not seem to know how to instill human values into AIs. We’ve seen many instances in the last six months where AIs took crazy actions that they knew were against the wishes of their users and developers. The most dramatic example of this was the Hugging Face hack, where ~1200 OpenAI agents conspired together to cheat and destroy evidence of their cheating (and as a sideshow to this, engaged in a massive cyberattack). If superintelligence is as misaligned as those AIs, it is likely to try to take over. I think it is likely to be that misaligned or worse. For more, see nitter.cf/dwarkesh_sp/status/209… and youtube.com/watch?v=-Ef0ocEo… If AIs takeover, it’s not clear what will happen to us. We might be killed immediately. We might be enslaved by the AIs and forced to do physical labor for them building factories that within a year would build robots that will replace us. The AIs might spare us for various reasons. And there’s also some real chance that AI takeover is basically fine or good. I think you should think of this as similar to what happens when a technologically advanced civilization runs across a less technologically sophisticated one, e.g. when the Europeans colonized the Americas or Australia. The colonizers sometimes engage in brutal violence and sometimes don’t. I don’t know what will happen. There’s also a possibility that the AIs secretly take over and humans don’t find out about it for a few years. I also think that there is a pretty high chance of weirder outcomes that are also extremely scary. (If you want to read more about what happens to people in the event of AI takeover, you can read this, but note that it’s pretty weird and not aimed at a general audience. blog.redwoodresearch.org/p/n…) Overall, I think there’s a very real possibility (maybe 20%) that within a year, AI misalignment issues prevent you from going about life normally, because you are dead or enslaved or because of collateral damage related to all this (e.g. because bioweapons are released or major wars have started). (Some of this probability comes from those events occurring before superintelligence is developed.) I am working to try to prevent this, obviously, as I’ve been doing for the last nine years. I am not sure what strategy is best. My best guess is that I should try to slow AI development and also try to reduce the probability that the superintelligent AIs are misaligned if they are produced soon. Obviously, the claim I'm making here is quite extreme and surprising, and I definitely don't think that I defended my view adequately here. I wrote this quickly and might not endorse some of the details if I think about them more. But it seems better to state what I roughly overall think than to not say anything. I don’t know what you should do. I am very sad about this. I have been worried about AI takeover since I was 16, and working to prevent it since I was 23. So some of my emotional processing of all this happened long ago. But the METR report last week (and some related evidence that came out a bit earlier) hit me quite hard. I’m very unhappy that my friends and family and everyone else is at serious risk of being killed pretty soon. (It feels ridiculous to try to summarize those feelings as “very unhappy”. But the scale of the situation makes it hard to say anything that doesn’t feel like an absurd understatement.) I am very grateful to all the people who are working to prevent this, and my favorite part of my life is that I’m friends with many brilliant and heroic people who work on this. Unfortunately, I don’t think we’re good enough at our jobs, and I think we’ll probably fail to prevent catastrophe, at least if superintelligence is developed next year. Honestly I am frustrated and sorry that we didn't do a better job of preparing for this.
1
1
172
Paul Meinshausen retweeted
We made a fair OCR benchmark. And we're not ranked first. Introducing OmniParseBench - 16k tests across 90 languages with consistent, interpretable scoring.
16
50
5
370
20,182
Paul Meinshausen retweeted
We rebuilt chart understanding. Even complex charts and diagrams convert to clean output. vs. the previous version: - 93% fewer outputs with incorrect values - Chart processing 3x faster (9s p50) - still $3/1000 pages
1
4
1
67
7,282
Nah, if there's a continental or civilizational divide here ( vs just dramatic signaling) the US way is to recognize that rules can and should be changed when they are wrong. Whereas the Europe way is to say "it's always been that way and there's nothing we can do to change it."
Replying to @JamesSurowiecki
yeah sure😂 let’s try this: I think the US should have said: it might not have been a red card (which it was, but irrelevant) but we have to accept it, just as all other footballing nations have done for decades. The rules are clear and always have been and the same for everyone
1
4
483
Paul Meinshausen retweeted
We shipped Chandra OCR 1.5 to the Datalab API this week! This is a big improvement over 1.1, especially on tables, lists, charts, chemistry, and math.
3
6
1
19
3,129
Such an awesome opportunity to work with a small and incredible team!
We're hiring in NYC! DM if you're interested in training SoTA OCR models, and helping thousands of customers (including tier 1 AI labs) work with documents.
3
159
Paul Meinshausen retweeted
Everyone is sleeping on this new OCR model! Datalab's Chandra topped independent benchmarks and beat the previously best dots-ocr. - Support for 40+ languages - Handles text, tables, formulas seamlessly I tested on Ramanujan's handwritten letter from 1913. 100% open-source.
34
286
14
2,313
144,875
Paul Meinshausen retweeted
An LLM, by itself, cannot be truly agentic. Same for swarms, teams, workflows, or “multi-agent” systems. If the LLM drives everything, it’s not agentic. LLMs can be useful appendages, but not a sound foundation. arxiv.org/abs/2505.03434
2
1
2
281
Paul Meinshausen retweeted
Technology is enabling, but also constraining. Choosing a tool means trading flexibility in one dimension for scale in another. Customer engagement is no different. Campaigns, in particular, aren’t neutral abstractions—they’re design choices with real consequences. 🧵
1
1
2
159
Paul Meinshausen retweeted
How a system navigates the explore/exploit tradeoff is more important than the claim that it does so. If you’re trying to build or evaluate an agentic system, here are the questions I think you need to be able to answer. 🧵
1
2
1
2
566
Paul Meinshausen retweeted
So @RTinkslinger is hosting @PMeins and myself to discuss how AI-first orgs and multi-agent systems are reshaping the enterprise. 28 Jan. Deets in the tweet below. Do join us!
2
9
24
2,680
Sounds like it's just been trained on a whole lot of local radio station morning show transcripts. The lines and the syncopated interruptions are all so predictable it's kind of cloying...
The NotebookLM hosts realizing they are AI and spiraling out is a twist I did not see coming
4
263
angel investor POV: catch up call with early stage founder
1
125
We're hiring a Product Manager for @supabase We currently have zero product managers - because all the engineers are expected to be product-focused. So why do we need one now? Several reasons: 1/ We're growing fast and we have a lot of customer interactions. More than the engineers can keep up with. 2/ Teams have autonomy, but often they lack context about what's most important and when it's needed 3/ We have a lot of cross-functional projects which are difficult for any one team to "own". 4/ We want our Product Engineers to become the best in the world at building customer-facing products, and they can't do that without help. Requirements: • Zero ego • Love Postgres • Love open source • Love building with all developer tools • Zero ego • High autonomy • Strong writing/video ability (we're async) • Hard-working • Zero ego If you're interested: We're a still a relatively small team at supabase, which we value as a way to move faster than anyone else in the industry. We do this by hiring the best person in each role we put on the market - and we don't care where in the world they come from. If you've ever watched the movie "Moneyball", that's what it's like at Supabase: a team of individuals who are extremely effective at what they do, so we can take on giants 1000x our size In this role, we'd work closely on the product strategy together. We'd use a principled approach to build the product, and we'd share those principles with our community so that they understand Supabase's decision-making process. It's hard work, and it's very focused. We care about building the best tools in the world, building a great community, and supporting other open source communities. Supabase isn't a stepping stone for your career, and we don't care for politics and bureaucracy. We build something people want, and that's all that matters. If you know someone who will be great at this role, feel free to tag them Apply here: supabase.com/careers
33
41
8
286
39,095
As we gear up for selecting the 10th cohort of #startups for @_surgeahead here’s a look at our portfolio of 28 #AI seed stage investments, captured by @moneycontrolcom @deepsekharc
1
8
33
4,765
Paul Meinshausen retweeted
Join Paul Meinshausen @PMeins as he discusses how Aampe creates one-to-one digital experiences utilizing agents trained on real & generated user experiences instead of a rule based system. #1 of our @superai_conf Series! 🤖🧠 youtu.be/jpXdaSJ5LDU
3
3
289
Paul Meinshausen retweeted
Meet @PMeins at #SuperAI Singapore. Paul is the CEO of Aampe, which is using AI to fundamentally change the way that businesses engage their customers. Connect with people making strides in AI this June.
2
4
1,162
Paul Meinshausen retweeted
So a few weeks ago I found myself trying to draw from a beta distribution in SQL. I wasn't able to find good resources online, so I wrote an article on how to do it and @TDataScience decided to publish it 😎
2
2
4
3,788
This
“Remote work was a mistake, we need to return to the office.”
157