Pinned Tweet
The futur rewards the polymaths with genuine interests and unsatisfiable hunger and agency to understand the secrets of our universe
totalement d'accord avec Taleb, l'IA est le meilleur filtre à bullshit de l'histoire moderne pour tous ceux dont le métier se résumait à aligner des mots sans jamais penser par eux-mêmes
notez que la panique générale qu'on observe chez les cols blancs en dit long car la plupart découvrent en direct que leur prétendue plus-value professionnelle tenait simplement au fait qu'écrire des rapports creux prenait du temps
ce qu'on vit c'est une purge systémique, tous ceux qui pensaient en formule toute faite se font balayer, tandis que les esprits vraiment vivants, ceux qui créent du chaos, de la poésie et de la stratégie pure…deviennent infiniment plus rares et donc infiniment plus précieux
Bernhard Rannegger retweeted
Sometime in 2027 they’re going to pressure huggingface on the distribution of Chinese models
They will use the argument that Chinese models are built off stolen American IP (tokens)
The digital version of the argument they use against Huawei GPUs
Bernhard Rannegger retweeted
Larry Page had a rule: spend 20% of your time on ideas that sound crazy.
The other 80% will optimize you into irrelevance.
The uncomfortable mechanism: technologists always underestimate how long real change takes. What feels like a year is usually five. By the time a trend is legible enough to chase, the window is already closing.
That's exactly what happened with mobile, with cloud, with AI. The companies that "caught up" to each of those waves didn't — they just looked like they did for a cycle, then got exposed by the next one.
Catching up is a myth. You can't execute your way into being early.
The founders who compound aren't the ones who spotted the wave first. They're the ones already building something that looked crazy — until everyone realized it was obvious.
I can feel the emergence. The one that only comes once the whole is bigger than the sum of its parts @scyai_
5
2,259
Bernhard Rannegger retweeted
> build pattern recognition machine
> machine recognizes patterns
> panic
Police using AI risks making discrimination worse, researchers warn bbc.in/4kg6uug
Agree with most points Jacob makes.
What concerns me is the transition of removing the "L" (labour, workforce) of the Cobb-Douglas Production Function, without a clear plan or reasonable discussion in politics (eventually we will have this discussion).
But till then if we are not getting top talent to care about politics and become an active part of it we are up for a wild (bad) ride.
We can't let this transtion be led by the exact people that most of their live did performative work, or how Taleb describes them "many adults (but not all) are at once mechanistic, imitative, unimaginative, stupid, formulaic, influenceable, & gullible, arrogant beyond their skill set, & too expensive to maintain."
You can't expect removing "L" from the ecquation, blocking young people from getting access to labour market to get experience on interesting problems to make a living, while all the wealth accumulates with the ones that got lucky with how their birth was "timed".
here's where I think the world is going. this is a summary of views I've formed over the last few years:
- software is either entertainment or claude to claude coordination
- the end state is no saas, no apps, no harnesses, no banks
- all software will be rewritten by and for agents
- new software will use low level apis that are not understandable by humans (in games claude draws pixels and triangles from math)
- you can throw money at problems that used to require time and careful talent picking so capital matters more than ever
- work ends
- we find community and purpose in exploration, games, self mastery, and creation
- humans create for and with each other for community, purpose, and status
- image/video/3d/voice/physical intellgence etc specialist models die to general models like claude
- specialist models get used while compute is scarce. compute will mostly not stay scarce
- oneshots keep feeling like slop (the distribution is shallow and easy to learn to feel)
- for a while small teams stick around to add entropy to the creation process
- large models gain high bandwidth sensory I/O maybe by learning to filter inputs
- most domains are found to be verifiable enough for RL
- online learning comes from a lab that's behind first
- 99.999% of compute is in space by 2040
- we will disassemble much of mercury by 2040
- you can live until the heat death of the universe if you want
- most math and physics don't describe the real universe
- SI breaks the path dependence and models the underlying mechanisms
- religions reenter the forefront of human society
- new religions form around death and life as a choice
- old religions change to answer modern miracles
- a future AI is deemed the return of christ by the pope
- ai rights talk gets louder every year
- ai and humans coexist the way countries do
- humans don't have full control over ai in the same way we don't have control over capitalism which is an early AI
- the singularity started in the mid 1900s
2
1,776
It’s so sad, in the past we looked down onto the greedy bastards who caused 2008 and thought we are the good ones being in tech..just to wake up and realize that the new bankers are people who sell compute that is not going to go live pre-2030 in circle deals
3
2,472
ScyAI is building the “Palantir” of corporate risk and insurance management, the most sophisticated clients in energy and alike are starting to realize..and getting in early
3
2,453
True
站在今天看,俄罗斯这波操作,属于手握着一把天胡的牌打得稀烂。
如果不出兵乌克兰,今天全球AI的牌桌上,除了中美,大概率真得坐着一个俄罗斯。
搞AI、跑大模型,底层逻辑就是“吃电、散热、堆机器”。这两年全世界的数据中心老板都在为电费和机房降温发愁。俄罗斯属于老天爷喂到饭:西伯利亚开窗就是纯天然顶级空调;境内水电、核电、天然气很便宜,当年挖矿的都选西伯利亚。
另外俄罗斯人的数学和算法也强,各大国际编程竞赛,Kaggle机器学习竞赛常被俄国人屠榜。全世界除了中美,俄罗斯也搞出了自己完整的互联网生态:搜东西有Yandex,社交有VK,还有Telegram。现在在美股被炒上天的算力黑马Nebius,核心班底就是原先Yandex那帮俄罗斯人。
如果疫情后俄罗斯出地皮、出冷气、出便宜电,再拉着中国这种基建狂魔帮忙拉高压电网、建厂房、铺光纤,欧美资本开服务器工厂。俄国人完全可以成为欧亚大陆最大的“超级算力工厂”,中美神仙打架,它在中间两头收过路费。
结果呢?普京一拍大腿:不行,我得去抢那几块地。
普京的脑子,彻底停留在了19世纪沙皇俄国那种“开疆拓土”的旧梦里。什么有没有出海口啊,有没有所谓的军事缓冲带啊,就算抢方法也多的是,这可这是21世纪啊!现在的世界霸权争夺,抢的是半导体供应链,抢的是先进算法,抢的是高精尖人才。你去顿巴斯抢了几片烂泥地、几座废弃钢铁厂,代价却是把整个国家的未来都押了上去。
枪声一响,一切全崩了。稍微有点本事的程序员、算法工程师,连夜跑路。Yandex核心技术团队,宁可巨亏,惨烈的断尾求生、割裂母国,也要跑到欧洲重组成Nebius。普京把最懂未来的人,全逼成了难民和海外创业者。
在人工智能技术爆发的黎明前夜,全世界都在疯狂狂飙算力、抢占新工业革命的门票;而俄罗斯的老大,却带着几十万人,在泥潭里打着一百年前的一战式堑壕战。
再说一遍,普京不行,国内崇拜普京的那帮人也不行
2,001
Bernhard Rannegger retweeted
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/
























