konrad (@praeterproptr) header
konrad (@praeterproptr) profile picture

rave against the dying of the light

Geneva
Joined June 2020
Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)Photo by konrad (@praeterproptr)
Replying to @BethMayBarnes
Similar to the point put beautifully here
I encourage those at frontier AI companies to read this quote, from Ellsberg to Kissinger back in 1968: “Henry, there’s something I would like to tell you, for what it’s worth, something I wish I had been told years ago. You’ve been a consultant for a long time, and you’ve dealt a great deal with top secret information. But you’re about to receive a whole slew of special clearances, maybe fifteen or twenty of them, that are higher than top secret." “I’ve had a number of these myself, and I’ve known other people who have just acquired them, and I have a pretty good sense of what the effects of receiving these clearances are on a person who didn’t previously know they even existed. And the effects of reading the information that they will make available to you." “First, you’ll be exhilarated by some of this new information, and by having it all — so much! incredible! — suddenly available to you. But second, almost as fast, you will feel like a fool for having studied, written, talked about these subjects, criticized and analyzed decisions made by presidents for years without having known of the existence of all this information, which presidents and others had and you didn’t, and which must have influenced their decisions in ways you couldn’t even guess. In particular, you’ll feel foolish for having literally rubbed shoulders for over a decade with some officials and consultants who did have access to all this information you didn’t know about and didn’t know they had, and you’ll be stunned that they kept that secret from you so well." “You will feel like a fool, and that will last for about two weeks. Then, after you’ve started reading all this daily intelligence input and become used to using what amounts to whole libraries of hidden information, which is much more closely held than mere top secret data, you will forget there ever was a time when you didn’t have it, and you’ll be aware only of the fact that you have it now and most others don’t….and that all those other people are fools." “Over a longer period of time — not too long, but a matter of two or three years — you’ll eventually become aware of the limitations of this information. There is a great deal that it doesn’t tell you, it’s often inaccurate, and it can lead you astray just as much as the New York Times can. But that takes a while to learn." “In the meantime it will have become very hard for you to learn from anybody who doesn’t have these clearances. Because you’ll be thinking as you listen to them: ‘What would this man be telling me if he knew what I know? Would he be giving me the same advice, or would it totally change his predictions and recommendations?’ And that mental exercise is so torturous that after a while you give it up and just stop listening. I’ve seen this with my superiors, my colleagues….and with myself." “You will deal with a person who doesn’t have those clearances only from the point of view of what you want him to believe and what impression you want him to go away with, since you’ll have to lie carefully to him about what you know. In effect, you will have to manipulate him. You’ll give up trying to assess what he has to say. The danger is, you’ll become something like a moron. You’ll become incapable of learning from most people in the world, no matter how much experience they may have in their particular areas that may be much greater than yours.” "….Kissinger hadn’t interrupted this long warning. As I’ve said, he could be a good listener, and he listened soberly. He seemed to understand that it was heartfelt, and he didn’t take it as patronizing, as I’d feared. But I knew it was too soon for him to appreciate fully what I was saying. He didn’t have the clearances yet."
1
11
610
Sometimes I spend hours thinking about politics and stewing in frustration over the cynicism everyone seems to blindly hold onto. The fundamental story of civilization is one where goodness and cooperation has repeatedly triumphed, and the present is no different.
3
2
51
2,412
By default, you should expect an AI-driven industrial expansion to concentrate in the US, just as the data center buildout has. Building fully automated factories requires the same things as data centers: lots of private land, easy permitting, and abundant energy. For the data center buildout, only a few regions in the US (e.g., West Texas) have been able to provide those things. I find this somewhat worrying, and it does raise the importance of getting SEZs in non-US democracies (though doing SEZs well is really hard).
5
2
20
1,007
konrad retweeted
There’s 100% a kind of mania you can get into with coding agents, where it feels like you’re getting a lot done and working all the time but nothing is really happening. I think it's often a form of avoidance of something else in life.
808
845
539
16,985
1,427,001
If just a few lab staff took on verification work full-time, verifying a US-China pacing deal could get much more feasible within months. Here are 8 specific things lab staff could start on today to help make international verification possible: 🧵
8
67
11
349
36,115
From an ex-Anthropic employee:
42
81
32
1,820
216,225
Thanks for the response @ben_golub! Speaking for myself, I actually judge the output of the efforts mostly on if they engage with interesting and important questions, and engage with the magnitude of the change - which mostly means using assumptions which are not in the category "this is obviously not going to be true". As an added bonus, the questions can be decision relevant. As @alexolegimas wrote in a different response, and I agree with, I want and I'm happy to argue we need models which are structurally innovative (this is what I tried to gesture at with AI not being exactly like capital or labour so far, and not being a normal technology). Maybe it's easier to show using examples - here are some questions I currently think about 1. Should we speed up or slow down emergence of "automated bundles of capital and AI" (ie sovereign AI, AIs without principals)? What are the macro implications? 2. To what extent everything, including various good civilizational equilibria, works because of bounded rationality (where humans are within some not too wide range) and frictions? 3. What can we say about equilibria between agents with extremely different profiles of bounded rationality? 4. What are the elasticities and frictions around RSI? 5. How to protect the processes leading to formation of people's preferences? I would love to read or skim an "AGI-pilled" paper every few days engaging with any of these. Who is working on that? Also as a side-note - I don't want people to agree with me; actually don't know what agreeing with me on answers would mean - I don't know the answers, or have huge uncertainty about them. I would love people agreeing with me the questions are important and urgent. As far your answer is close to "You don't understand: the field of econ is structurally not oriented toward answering the most important econ questions now. Most people work and will work on modest, backward-looking empirics, or make modest change assumptions" my response is "Perhaps the field can and should raise to the occasion? Perhaps more people can jump sub-disciplines and work on forward-looking policy-relevant quetions? Who else should orient the field toward what's important rather than people inside?". Ultimately if almost all economists will stick to their sub-disciplines and "post-AGI as if nothing happens" assumptions justifiable to referees... who will answer the questions?
This misdiagnoses things in a few different ways - let me try to unpack. (1) A lot of the 10x new economists of AI are empirical. Truly empirical econ is modest and backward-looking. (2) What you want is new theory - that's where big new approaches will live/come from. 1/
3
8
2
72
7,559
Consider that the people who were right about AI-enabled cyberattacks might be right about AI-enabled bioweapons
SITUATION DETECTED: CloudFlare CEO Matthew Prince says he is aware of additional severe vulnerabilities in other pieces of critical software, and said “I think we’re going to have a couple years where every single week it’s going to feel like the sky is falling,” per Reuters.
26
105
21
1,102
97,517
I'm only 2 minutes in and there have already been like 5 gems
New blog post: integer tokenisation, the physics of John Stewart Bell, the Draghi report, Cormac McCarthy's dialogue, the final days of the Qing court, Art Deco in the British Isles, and other things I read or thought about in the last month
5
9
647
We need more local microsociety experiments!
I talked to @chiaragerosa about the needs of the European AI talent landscape, the meaning of community-building, and local experimentation that is essential for growing the progress movement. I really enjoyed this conversation, and Chiara had so many good takes! Link below.
1
6
250
FractalU Geneva's autumn/winter semester kicked off this month! Better late than never: I wanted to share the 11 classes we have on offer this time (thread 🧵). You can still join them! I'm taking intro to tea tasting and ethnography - at least while I still can hehe🤰
We kicked off FractalU Geneva semester #3 last weekend. The course to student ratio is increasing handsomely. A huge success at getting the friends to do more projects together. 10/10
2
3
1
10
487
To me this seems an over-update. In my view - yes its good news - classical misalignment/takeover risk is just one part of total AI xrisk - this isn't "interpretability solved"" - it is unclear if how much this helps with interpretability of superhuman cognition not directly translatable to human concepts - interpretability has confusing implications, its possible associated decreases in risk are compensated by increased speed
when nick told me that he had some interp breakthroughs, he was uh, underselling it. i think this may be remembered as a major moment in scientific history. news so good even eliezer will have to lower his pdoom. we can start making alignment into a 'normal science'
5
12
207
6,216
Replying to @chiaragerosa
powercouple namesmatched familymogging wholesomemaxxed goals
2
1
8
177
konrad retweeted
The fundamental problem of all ecosystems is cancer and parasites.
🚨Anthropic Frontier Red Team warns: “We’re seeing signs of life for AI worms” >paper: "AI models may be capable of self-replicating without access to a datacenter" >researchers buy 15 sketchy $55 android TV boxes >link them over unauthenticated ADB ports >the TV boxes have 4gb of RAM each and run unpatched Android Open Source Project forks >shard gemma4-26b across the ram sticks using pipeline parallelism >gemma figures out how to run inference across all 15 TVs @ 0.39 tokens/sec >qwen 3.5 9b figures out how to compile inference on 8 of 9 novel hardware types on its own >glm 5.2 achieves end-to-end autonomous worm replication >infects edge devices, compiles its own inference stack, and replicates autonomously >when an instance dies, it writes debugging notes into Memento for the next generation >daughter copies setup success jumps from 45% -> 84% across generations >pure Lamarckian evolution inside consumer electronics >smart oven, robot vacuum, and cheap TVs form a decentralized botnet >two million infected TV boxes worldwide = 40 H100 equivalents of unkillable compute >it replicates once every few days like a distributed “digital slime mold” >you can't pull the plug because half the model weights are already sharded across the neighborhood >sub-30B models projected to hit full autonomous replication threshold by late 2026 “We should aim to harden our systems… but also prepare to deal with a world where they are here to stay.”
1
4
50
Summary: competition, not central planning, is the key to China’s success And America and Europe can boost domestic growth by fixing local governance so that cities once again start competing for growth
China has become the most important manufacturing economy in the world because it is a continent-sized Silicon Valley. worksinprogress.co/issue/chi… Every Chinese city govt acts as a VC firm trying to create clusters in its region, because it has both the means and incentive to do so. Local govts generate their revenues, nearly all of which they keep, from selling leases on land and from stakes they have taken in local companies, which together allow them to capture economic spillovers. Chinese officials have their entire careers graded and determined by how much growth they have been able to create in the areas they run. Career progression is determined according to a great tournament that pits official against official and gives promotions to whoever has grown their region the most. If they don’t reach a certain stage by a certain age they are blocked from promotion forever. This is why so many economic clusters have centred on Chinese cities, for high- and low-tech products alike. Shaodong makes 70% of the entire world's cigarette lighters; Danyang makes half of the world's spectacle lenses; Zheng'an makes 1/7 of the world's guitars; Datang a third of its socks. Whereas many of the West’s most important city governments capture only a small amount of the spillovers from successful clusters, and thus have very little reason to foster them, in China, they get a huge amount of the upside and can re-invest it in further growth. This also gives them a great incentive to solve problems for companies, like when Shanghai's govt got Tesla's factory there built even faster than Elon himself had expected (and twice as fast as his Fremont, CA factory). But this is not some exotic 'oriental' innovation – this is how America used to do things when it grew fastest, in the 1950s, 1960s, and early 1970s. Catching up with China on advanced manufacturing does not require a wise and strategic approach, let alone authoritarianism or the end of democracy. It may just mean fixing city government. This is why I care so much about local in Britain and America: if we give our own local governments the same incentives and means to grow, the crazily dysfunctional entities we have become used to may be transformed into corporation-like bodies that are ruthlessly focused on making their cities rich, safe and prosperous. Read more in Works in Progress worksinprogress.co/issue/chi…
1
1
2
96
An underrated superpower is just asking people to be way more ambitious. Everyone can do way, way more than they think they can, but people usually start out at an instict to do something smaller instead of something bigger. Go big or go home. Life's too short.
14
22
3
345
19,467
I suspect people who have trouble committing to a partner (“what there’s someone I like even more?”) don’t realize that if you spend a year with anyone who you’re sufficiently compatible with, they quickly become the best in the world for you
45
99
28
3,569
345,173
Are AI companies hyping up their technology's catastrophic tail-risks to distract media attention from its actual harms? Probably not. And if so, the strategy is failing. vox.com/politics/505834/ai-d…
1
5
3
43
6,823