Greg Janik
4.3K posts

Greg Janik
@gpjanik
wandering around SF & building
San Francisco, CA Katılım Şubat 2013
575 Takip Edilen262 Takipçiler

Back in 2014 I was on a walking tour of Dresden and our guide kept referring to the firebombing destroying this or that building until an American girl in her 20s raised her hand to ask "sorry, who firebombed Dresden?" and the group fell so silent you could hear a pin drop
Matthew Yglesias@mattyglesias
Dresden was rebuilt surprisingly recently
English

@NikTek looks million times better than the flat shadowless crap
you can do color grading and artistic direction differently with dlss it just happens so that most games didn’t
English

I can't believe that Nvidia looked at this "AI on top of games filter" and said to themselves this is the future of gaming.
Like it or not, this is where Nvidia is heading and they're calling it neural rendering with DLSS 5.
The examples they've showed reminded me a lot of those AI generated filter videos on top of GTA V, except that now it is supposed to run in real-time.
Honestly, I don't like this current look at all
English

@ezoteryczny To na zdjęciu jest git. Jest dobra kolorystyka, dobra wysokość/skala, budynki są w pierzei a nie rozrzucone jak gówno po polu. Odległości między oknami normalne, jak na przeciętnej ulicy.
Polski

@BReguided bc it looked like tier 3 office building in Gary Indiana
English

@damian_patecki warszawa to jest najbrzydsze duże miasto w europie i długo zajmie, zanim przestanie być, sorry
Polski

@garrytan mostly by the lack of demand
startup scene has consistently terrible taste
English


@lacherbauer tokyo has roughly as many people as spain, 200 miles is basically nothing lol
English

We believe Cursor discovered a novel solution to Problem Six of the First Proof challenge, a set of math research problems that approximate the work of Stanford, MIT, Berkeley academics. Cursor's solution yields stronger results than the official, human-written solution.
Notably, we used the same harness that built a browser from scratch a few weeks ago. It ran fully autonomously, without nudging or hints, for four days.
This suggests that our technique for scaling agent coordination might generalize beyond coding.
English

@PeteHegseth @AnthropicAI @DarioAmodei @POTUS you and your boss are now smaller than the third biggest AI company, which is funny
English

Greg Janik retweetledi

Polish $58k GDP vs Spanish $58k GDP


Visegrád 24@visegrad24
Poland has now passed Spain in GDP per capita in PPP The latest figures from the IMF: 🇵🇱 Poland - $58 559 🇪🇸 Spain - $58 348
English

@sikorskiradek A co dokładnie wg Pana w temacie broni atomowej jest niepoważne?
Polski

@burkov “The processing of input is done in parallel by hardware anyway”
this is about as smart as “money is just an entry in the database you can just add more and people will be rich”
English

LLMs process text from left to right — each token can only look back at what came before it, never forward. This means that when you write a long prompt with context at the beginning and a question at the end, the model answers the question having "seen" the context, but the context tokens were generated without any awareness of what question was coming. This asymmetry is a basic structural property of how these models work.
The paper asks what happens if you just send the prompt twice in a row, so that every part of the input gets a second pass where it can attend to every other part. The answer is that accuracy goes up across seven different benchmarks and seven different models (from the Gemini, ChatGPT, Claude, and DeepSeek series of LLMs), with no increase in the length of the model's output and no meaningful increase in response time — because processing the input is done in parallel by the hardware anyway.
There are no new losses to compute, no finetuning, no clever prompt engineering beyond the repetition itself.
The gap between this technique and doing nothing is sometimes small, sometimes large (one model went from 21% to 97% on a task involving finding a name in a list). If you are thinking about how to get better results from these models without paying for longer outputs or slower responses, that's a fairly concrete and low-effort finding.
Read with AI tutor: chapterpal.com/s/1b15378b/pro…
Get the PDF: arxiv.org/pdf/2512.14982

English

@MarcinMozdzonek z tego samego powodu z którego mają 2x wyższą płacę średnią
Polski

Ciekawe dlaczego Niemcy nie biorą udziału w programie SAFE?
Donald Tusk@donaldtusk
Maski opadły. PiS i Konfederacja zagłosowały przeciwko programowi SAFE, czyli przeciw bezpieczeństwu Polski, przeciw nowoczesnej armii, przeciw polskiemu przemysłowi zbrojeniowemu. To już nie jest opozycja, to są wrogowie polskiej niepodległości.
Polski

@shanggyangg it is way closer culturally to luxembourg than to belarus in every possible sense
English

There is simply no world in which Lithuania is culturally closer to Luxembourg than to Belarus
Valus Mäng@Viktor_Twittor
They're calling it "The Pilsner-Belt"
English
Greg Janik retweetledi

Sport shouldn’t mean amnesia, and the Olympic movement should help stop wars, not play into the hands of aggressors. Unfortunately, the decision of the International Olympic Committee to disqualify Ukrainian skeleton racer Vladyslav Heraskevych says otherwise. This is certainly not about the principles of Olympism, which are founded on fairness and the support of peace.
I thank our athlete for his clear stance. His helmet, bearing the portraits of fallen Ukrainian athletes, is about honour and remembrance. It is a reminder to the whole world of what Russian aggression is and the cost of fighting for independence. And in this, no rule has been broken.
It is Russia that constantly violates Olympic principles, using the period of the Olympic Games to wage war. In 2008, it was the war against Georgia; in 2014 – the occupation of Crimea; in 2022 – the full-scale invasion of Ukraine. And now, in 2026, despite repeated calls for a ceasefire during the Winter Olympics, Russia shows complete disregard, increasing missile and drone strikes on our energy infrastructure and our people.
660 Ukrainian athletes and coaches have been killed by Russia since the full-scale invasion began. Hundreds of our athletes will never again be able to take part in the Olympic Games or any other international competitions. And yet, 13 Russians are currently in Italy competing at the Olympics. They compete under “neutral” flags at the Games, while in real life publicly supporting Russian aggression against Ukraine and the occupation of our territories. And they are the ones who deserve disqualification.
We are proud of Vladyslav and of what he did. Having courage is worth more than any medal.

English

🚨IOC statement 1/3
Skeleton pilot Vladylsav Heraskevych not allowed to participate at Milano Cortina 2026 after refusing to adhere to the IOC athlete expression guidelines
Having been given one final opportunity, skeleton pilot Vladylsav Heraskevych from Ukraine will not be able to start his race at the Milano Cortina 2026 Olympic Winter Games this morning. The decision followed his refusal to comply with the IOC’s Guidelines on Athlete Expression. It was taken by the jury of the International Bobsleigh and Skeleton Federation (IBSF) based on the fact that the helmet he intended to wear was not compliant with the rules.
The International Olympic Committee (IOC) has therefore decided with regret to withdraw his accreditation for the Milano Cortina 2026 Games.
Despite multiple exchanges and in-person meetings between the IOC and Mr Heraskevych, the last one this morning with IOC President Kirsty Coventry, he did not consider any form of compromise.
The IOC was very keen for Mr Heraskevych to compete. This is why the IOC sat down with him to look for the most respectful way to address his desire to remember his fellow athletes who have lost their lives following Russia’s invasion of Ukraine. The essence of this case is not about the message, it is about where he wanted to express it.
English















