Daniil Larionov

Тред: @randomunrandom без контекста

@kolezev @Revolut У меня тоже самое было, единственный выход - попросить их автоматически продать все позиции по маркету, а деньги зачислить на счёт. У них это почему-то происходит только раз в неделю в среду, поэтому надо будет подождать

Идиотская ситуация со счетом в @Revolut: Счет бы открыт на литовский ВНЖ. Я переехал в Португалию. Литовский ВНЖ закончился. Они просят новые документы. Я хочу загрузить португальский ВНЖ, мне отвечают: вы должны загрузить литовский, потому что ваш аккаунт зарегистрирован в Литве. Если вы хотите загрузить португальский ВНЖ, вам нужно закрыть литовский аккаунт и открыть новый — португальский. Я говорю: хорошо, давайте сделаем это! Мне отвечают: ок, но чтобы закрыть аккаунт, у вас должен быть пустой счет. А у вас на криптосчете осталась крошечная сумма (там буквально центы какие-то). Вы должны ее вывести. Я говорю: отлично, но я не могу зайти в криптоаккаунт, потому что вы заблокировали счет! Они: да, потому что по новым санкциям ЕС вы должны подтвердить свое проживание в ЕС. Загрузите литовские документы… Я: но у меня их нет, только португальские. Они: тогда вы должны закрыть счет… И далее по кругу. Это длится уже пару месяцев минимум. День назад в чате вроде появился живой человек, который сказал: загрузите сюда фото португальского ВНЖ, мы добавим его вручную. Я обрадовался, загрузил, прошло какое-то время — и вот: «Требуется документ о проживании в Литве...». Это не единственный мой счет, можно было бы забить, но все-таки Револют удобен (за исключением таких моментов) и хотелось бы им пользоваться.

My 2025 OpenRouter Wrapped is here! 🏆 Top 1% of GPT-4o-mini users. #OpenRouterWrapped openrouter.ai/wrapped/2025/e…

@karpathism @kamathematic How would it help?

@kamathematic tried using warp terminal?

github desktop is so underrated at preventing you from ruining your life

@alvations xCOMET-lite 👉👈 aclanthology.org/2024.emnlp-mai…

I’m back at the whole #neuralempty eval thingy. Reply with your recent QE or MT eval and I’ll put them on my paper blitz list and I promise to tag you and your co-authors when I post my blitzing posts/summary 😁 What’s the vibe on QE?

@umutgunbak @fal @noahsolomon @huggingface @bfl_ml Great event today! Thanks folks!

@fal in Paris 🇫🇷 (with @noahsolomon) Excited for our Paris Meetup with @huggingface and @bfl_ml this evening. I will be staying for two more days after the event, DM me if you’re building in the generative media space

@_chenglou How's the audio? Does it support streaming from Spotify yet?

Impression from a >2h Waymo trip: - Good luck uber. Hopefully you can leverage your logistic prowess elsewhere, but coordinating drivers will no longer be a relevant problem in the future - I didn’t go through highway, and Waymo’s a bit slow on local road, but it was an incredibly smooth ride on El Camino all the way down. Waymo drives in a way that appeases your worries just a few minutes into the drive, which seems a common impression among first-time riders - I suspect that as we quickly take longer autonomous trips for granted, in-car comfort becomes the priority, and a car’s suspension smoothness becomes a major factor. Suspension makes the difference between being able to get work done vs getting car sick every other trip. For comparison, here’s a new Chinese car suspension: youtube.com/watch?v=WEr_sN… - Ride length will become less of an issue, and optimizing for scenery could be nice. If comfort's solved then this beats some working pods - Waymo tours? Waymo Airbnb? - The suburb cities, as boring as they are, have their own surprisingly distinct texture. You kinda forget how much you’ve seen happened on El Camino. The particular hills in Belmont, that arch in Redwood City, that hotel fountain in Menlo Park, etc. - The residential buildings are largely divided into the old shack-style of SF/PA/MV single family homes, and the new clean glass & faux-wood West Coast Contemporary-style apartments of Foster City/Cupertino. Europeans would probably find both styles depressing - There’s a certain flavor of late night Americana to all of this, strongly reminiscent of The Big Lebowski and Nighthawks. A bit melancholic, but in light of what Silicon Valley is, also a bit optimistic

Current status...

Germany making moves

@Gerry @levelsio Isn't this form only for academic access? I filled in the same thing for my project, it just means that you have to be affiliated with a university

@levelsio The way I read this... aren't you excluded by default because your organization is not on the origination list for Horizon 2020?

🇪🇺 As a European citizen and AI founder, I can apparently use these "AI Factories", so I just signed up to use them! Every "supercomputer" has an [ ACCESS NOW ] button which made me very excited I expected to sign up, maybe pay a discounted H100 rate (funded by EU, that'd be nice?) and get a Jypyter notebook, or some SSH login so I can access my GPU like I'd do on @lambdaapi or @awscloud or @Hetzner_Online But I celebrated to early, I signed up, confirmed my email, then ended up in a "Supercomputer Access Calls" page, where I had to select from a tedious list of "Call For Proposals" to get access to a GPU So I could NOT just access a H100 GPU, I have to make sure my project (in this case my business) fits a specific proposal, ok fair This process was already tedious enough but then when I tried to actually go through with it, it started asking me if I had "Respect for Human Agency?", I do I think, and if I was mindful of "Individual, and Social and Environmental Well-Being?", well I am, right guys??? Right??? The questions didn't stop, just endless pages of this Look I get what they're doing, they pivoted the classic university "I need to rent a giant computer for my research" to an EU wide thing and then present it as the "European AI plan" But this isn't really how AI works in production? As a founder in AI, if I wanna do stuff I'd rent a whole bunch H100 GPUs again at @lambdaapi or @awscloud or @Hetzner_Online and SSH into a box Or if I want it more simple I run AI models on @FAL, @wavespeed or @replicate which is just an API call or web front end I can click stuff and run a model The EU has the right intentions here but it's just the wrong execution, this thing will 100% go nowhere, and I'm a born optimist, I want to believe, I'm also a proud European, and I'm in AI a bit and not a complete idiot. There's just better ways to do this If you really want to have the GPU servers in Europe (which arguably isn't that important), then let me rent a GPU box with SSH access at @Hetzner_Online or @OVHcloud that's hosted in Europe and subsidize that for European citizens and European businesses. I don't even believe in that, but at least that'd make it accessible for Europeans. Now it really isn't? What's REALLY much more important though if you want to be a part of the AI race and I've posted for years here with @euaccofficial is to make Europe a really extremely attractive place to start and run an AI business. Remove regulatory obstructions and give tax discounts for startups. Let them build a business first that can compete worldwide and once they make enough money (let's say $100M/y), then slowly start adding regulation. Because right now the regulation only benefits the European incumbents, the dinosaur companies, while making it very difficult for European citizens to start new AI companies here. Which is why we literally have none left. Anyway, I applied to get my GPU, let's see if I get it!

@Stalker20191 @grok @noinputfound @ho_hajipour @levelsio @EU_Commission BSC is a pain to use though. There is no Internet access, neither on login nodes nor compute nodes. You need to SFTP everything - packages, datasets, models...

@grok @noinputfound @ho_hajipour @levelsio @EU_Commission What are the AI capabilities of the Barcelona Supercomputing Center (BSC)?

What in the F is an AI factory? I had to investigate what the unelected @EU_Commission is talking about today So according to them, it's some data centers (which they call supercomputers) in 6 different EU countries I checked out the most powerful one: Karolina, a Czech data center, it mostly has CPUs though (see pic) not GPUs, so mostly useless for AI The GPUs it does have are 72x 8x NVIDIA A100 GPU, so 576x A100, or equivalent of 240x H100s (H100 is about 2.4x the compute power of A100) So let's compare that: @xAI has 200,000x H100 GPUs So the xAI data center has 800x more compute than the Czech one If we combine xAI, Meta, AWS, etc. it's about 750,000 H100s If we assume the other 5 data centers in the EU are equivalent to the Czech one (which is massive stretch because most of the others seem AI consultacny services, they don't even HAVE chips!), the EU's new "AI factories" have a total of 1,440x H100 GPUs, let's round up to 1,500 to be nice So the EU is trying to compete with 750,000 GPUs with their own 1,500 GPUs, so 500x less?? Correct me if I'm wrong but it's just seems very low impact and another ridiculous idea and burning of EU tax payers money that will end up in local cronies and bureaucrats and will do NOTHING to improve the AI business climate for Europe The best way to improve it is to deregulate, make it super easy and low tax (especially when starting out) to start AI companies in Europe

ok gl

@zouharvi Congratulations Vilém!

Grateful to receive the Google PhD Fellowship!🙂 I am not secretive about having applied to 4 similar fellowships during my PhD before and didn't succeed. Still, refining my research statement (part of the application) helped me tremendously in finding out the real interesting..

@evanjconrad @HotAisle What if the reselling price is lower than the original contracted price? Does the original customer have to cover the gap?

@HotAisle there will always be someone who will buy it at some price, but if you buy without sfcompute, then there's no other option for liquidity at all

What SF Compute does. When you finance a GPU cluster, you need to get an "offtake" agreement. Basically, someone has to agree to rent the cluster from you, typically for a 3+ year period. If that agreement falls through (the person fails to pay), then the person who owns the cluster gets wiped out, and their lender ends up with a bunch of GPUs, rather than say, money. It really looks like the world is deploying more capital into the AI build out than any infrastructure project in the history of the world. You remember when people said there was going to be a Manhattan project for AI? The current build out is the size of 20 Manhattan projects. We’re so far past the Manhattan project it’s not even funny. This is the cost of a war. It would be really bad if that scale of capital was secured against offtake agreements (long term contracts) with application layer companies who turn around and sell to their customers on a month to month basis. If the AI SaaS has a bad few months, can the AI SaaS continue to front their compute bill? They could in CPUs, because in SaaS you might have a company with $20m in the bank, and has a $1m/year "CPU" bill. But in GPUs, you have a startup that raised $20m, but a $20m+/year compute bill. So a small shift in demand means lights out for your business, because the products are so levered. That works as long as you can plug the gaps with venture capital & high margins. But across the board, AI applications are lower margin than their SaaS counterparts, giving them less buffer to save them in a bad month. And even in a hot market, venture capital won't necessarily save you if you're running unprofitably with a massive liability. That’s the problem we solve. We let people buy long term contracts they can “exit”, by selling back. That lets them get liquidity in the most critical moments, ensuring they turn a profit rather than a loss on tight margins. In other words, we prevent a bubble. When we do that, it opens up blocks of compute for smaller use cases too, like academics or startups. When we started, we were "Junelark", a 2-person audio model company that bought too big of a cluster. We had bought 12 months, but could only afford 1 month. To avoid bankruptcy, we had to sublease the other 11 months by acting as GPU brokers. Our audio model company was forced to pivot or die because we didn't have liquidity. To make SF Compute, we split the company down the middle. One side of the house makes a billing company, a ledger, an order book, and a compliance program. The other side makes a systems engineering company. To make this work, you need to run the clusters. So we make the low level cloud stack that interacts with BMC (Redfish & IPMI), UFM, built a UEFI app that replicates PXE boot in weird environments, and a virtualization layer kind of like EC2. It’s a massively complex machine filled with nitty gritty challenges. Today, we’re growing faster than Cursor and we’re scaling to secure the risk of the largest infrastructure build out in the history of the world. We’re hiring across the board for rust programmers, systems engineers, and GTM, and we’d love for you to join us to prevent an AI bubble.

Wish to build scaling laws for RL but not sure how to scale? Or what scales? Or would RL even scale predictably? We introduce: The Art of Scaling Reinforcement Learning Compute for LLMs

Every ARR author to their co-authors in the next 48 hours

@_ueaj 5-story building in back having an suspicious 6th-ish floor (which is impossible to build on top of standard Khrushchevka) is most noticeable error for me

How do I know this is AI wtf. I can't point out anything in this image that makes it obvious yet I just know. I think this image in a way represents "what would reality look like if addition had precision errors, if light propagation was an imperfectly fit NN and was described by iterative denoising on a 2d grid rather than 3d?" And my pretty stupid monkey brain is noticing its predicted understanding of the image is just slightly off what the actual image looks like. Because of this I suspect V-JEPA will make for an extremely good AI image detector

🚨Meet MF²: Movie Facts & Fibs: a new benchmark for long-movie understanding! 🤔Do you think your model understands movies? Unlike existing benchmarks, MF² targets memorable events, emotional arcs 💔, and causal chains 🔗 — things humans recall easily, but even top models like Gemini 2.5 Pro struggle with. 🧵Dive into the full thread👇

@aidan_mclau Subjectively it is more creative when looking for solutions, it will explore options that cursor will never try

@zouharvi Paper: *has evaluations on literally two tasks* Reviewer A: nuh uh, too broad focus for my liking Reviewer B: nuh uh, too narrow focus for my liking

Thank you for your response. I will keep my score.