Pinned Tweet
In the future they will make an AI version of you and train it on all your tweets. This is probably not true, but you should tweet like it is.
That policy is not designed to tackle housing shortage. It's designed to express anger at the landlords, or something like that. Unfortunately the cost of such expression is quite high.
Good Morning from Germany, where one of Europe’s biggest landlords calls the country “uninvestable.” Heimstaden is withholding fresh capital over Berlin’s expropriation plans. It owns 20,000 apartments in the city worth ~€5bn and estimates a potential €1bn loss. Its hybrid bond has plunged from ~98 cents in August to 82.5. The irony: a policy meant to tackle the housing shortage risks driving away investment without building a single new home. bloomberg.com/news/articles/…
I agree, but it's not so binary. Parts of modern human civilisation are designed, it's not natural selection alone that shapes us. Likewise, natural selection plays a role in creating AI. For example, AI products that win in the market are more likely to be copied.
Henry Shevlin (@dioscuri), Dan Williams (@danwilliamsphil), and I discuss AI and "superintelligence."
You get all of these maladaptive outcomes—all this cruelty and waste and competition and aggression—precisely because there was no designer who wanted to minimize any of these things. Natural selection did its thing.
If AI were simply allowed to evolve without any kind of human intervention, then that would be a risk. But in the case of AI, there is intelligent design. I mean, us. These are gadgets. I think it’s illicit to anthropomorphize AI as being like human beings, which are products of natural selection, when AIs are not products of natural selection. They’re products of intelligent design.
And so the engineers can think, “Well, these are the goals that we want to optimize. These are the kinds of side effects that we want to minimize. These are the kinds of things that we wouldn’t in a million years allow it to do.”
Vasily Kuznetsov retweeted
Here's a 45-second TL;DR on Jev.
I find the core idea beautifully simple, but the video made it really hard to understand.
Hope you find it helpful.
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
Here's a calculator so y'all can be sure how to vote in the next red/blue poll, based on your values and assumptions: entwine.info/red-blue-poll.h…
If my grandmother had wheels, she'd have been a bicycle
NEW: Elon Musk took $500 million in loans out at SpaceX, a move that would have been illegal at a public company. It's just one example of the years of financial engineering at the helm of his companies. Latest investigation with @susannecraig nytimes.com/2026/04/24/techn…
Readers added context they thought people might want to know
Musk's ~$500M SpaceX loans (2018-2021) at low rates were repaid in full by end-2021 with ~$14M interest. Legal and routine for private company owners; Sarbanes-Oxley prohibition applies only to public firms.
nytimes.com/2026/04/24/tec…
lw.com/admin/upload/S…
Vasily Kuznetsov retweeted
Regardless of what happens in Europe and the U.S., the world is moving to EVs.
There is nothing that western automakers can do to reverse this trend, and geopolitical risk in oil will only accelerate it.
Either get good at EVs, or wither away.
Emerging Asia is sprinting toward electric independence to crush fossil fuel fragility. Fossil fuel wars will only accelerate this trend
➡️Viet Nam: 38% EV sales share, ahead of EU
➡️Singapore: Leading the charge at 50%+ share
➡️Indonesia: 15% share, more than US
➡️Thailand: 21% share, leaving Japan (at 3%) in the rearview mirror
➡️India: Three-wheelers at 57% share, EVs rising rapidly and at 4%, already ahead of Japan
➡️China: Already banking $28b/year in avoided oil imports
The era of being held hostage by oil chokepoints is coming to an end in multiple Asian countries, fast
ember-energy.org/latest-insi…
The Patel brothers dropped another episode 🔥
.@dylan522p gives a deep dive on the 3 big bottlenecks to scaling AI compute: logic, memory, and power.
And walks through the economics of labs, hyperscalers, foundries, and fab equipment manufacturers.
Learned a ton about every single level of the stack.
0:00:00 – Why an H100 is worth more today than 3 years ago
0:24:52 – Nvidia secured TSMC allocation early; Google is getting squeezed
0:34:34 – ASML will be the #1 constraint for AI compute scaling by 2030
0:56:06 – Can’t we just use TSMC’s older fabs?
1:05:56 – When will China outscale the West in semis?
1:16:20 – The enormous incoming memory crunch
1:42:53 – Scaling power in the US will not be a problem
1:55:03 – Space GPUs aren't happening this decade
2:14:26 – Why aren’t more hedge funds making the AGI trade?
2:18:49 – Will TSMC kick Apple out from N2?
2:24:35 – Robots and Taiwan risk
Look up Dwarkesh Podcast on YouTube, Apple Podcasts, or Spotify. Enjoy!
Vasily Kuznetsov retweeted
Love to see what @RichardMCNgo is doing here in collaboration with @SamoBurja.
That I have read the majority of the reading list independently already shows how the discourse currently converges in important ways.
If you haven't read these resources, consider signing up for working through this 21st Century Civilization Curriculum.
Also consider it, if you already have. I just did.
Many aspects of Western civilization are far less functional than they used to be. What went wrong, and how should we respond?
My new curriculum (drawn up in collaboration with @SamoBurja) aims to answer those questions.
Read it and enroll in discussion groups below.
In our timeline Butlerian Jihad is called Yudkowskian Shabbat.
@ESYudkowsky @PauseAI
Vasily Kuznetsov retweeted
Today we're announcing Gauss, our first autoformalization agent that just completed Terry Tao & Alex Kontorovich's Strong Prime Number Theorem project in 3 weeks—an effort that took human experts 18+ months of partial progress.
A talk about AI progress that doesn't seem out of date 13 years later. Also still funny :D
Amazing that @SchmidhuberAI gave this talk back in 2012, months before AlexNet paper was published.
In 2012, many things he discussed, people just considered to be funny and a joke, but the same talk now would be considered at the center of AI debate and controversy.
Full talk:
How did @TIME select people for #TIME100AI that they don't even have @demishassabis there? These lists are often so puzzling.
TIL that indicators can be not only green/yellow/red but watermelon: green on the outside, red on the inside. The aggregated data look good, but when you disaggregate you see problems.
For example: average/median salary is growing, but in the lower deciles it's going down. Or overall sales are growing, but some (important) segments are collapsing.
true story: scienceblog.com/universe-set…
