Pinned Tweet
Welcome to 𝕏.com terminal.
Please log in and enjoy your stay.
English
Error: Failed to read "username"
84 posts
Error: Failed to read "username"
@errors_here
Access violation at address 00000000 in module 'twitter.dll'. Read of address 00000000.
Joined Ocak 2023
100 Following9 Followers
Nie mówiąc gdzie pracujesz, napisz jedno zdanie, które powtarzasz 100 razy dziennie, reszta zgaduje jaki zwód💁♀️
Polski
@byteHumi Needed for reinforcement learning
English
@FakePsyho Do you have any rate limiter issues?
English
I got tired of engagement-farming AI reply bots, and a few hours later I had a firefox extension that correctly tags (and hides/collapses if enabled) about 90% of them.
Turns out you can detect most of these accounts just by analyzing their posting patterns. Take their last 200 tweets and find the maximum number of reply tweets within any 5-minute window. If it's 4 or higher, it's an automated reply account. I manually tagged 100 suspicious accounts and 50 terminally online accounts. This single if matched 86/100 of sus accounts and none of the terminally online ones.
I've been running this for a few days now and I'm still surprised by the accuracy. I'm pretty sure you can detect almost all of this AI slop without even looking at the content of tweets. I manually checked "false positives" I got and all of them were either: (a) accounts spamming gifs/single words/emoji replies, which is more or less the same behavior just without AI (b) accounts that used AI replies in the past.
tl;dr: fixing majority of the current slop apocalypse requires a 0.1x engineer and a minimal effort
English
@andrzejdragan Te benchmarki mają testować inteligencję "ogólną", jak szybko modele się odnajdą i zrozumieją jak wykonać zadanie dotąd nieznane i bez wyjaśnienia jaki jest cel tych gier, po prostu próbując o swoich siłach co można tu zrobić i jak to działa.
Polski
Do AGI jeszcze sporo nam brakuje, ale testowanie modeli *językowych*, których mocną stroną miało być operowanie na tekście, przy pomocy gier wideo, nie różni się nadmiernie od bicia dziecka po głowie, bo nie widzi w podczerwieni.
Guri Singh@heygurisingh
Humans: 100% Gemini 3.1 Pro: 0.37% GPT 5.4: 0.26% Opus 4.6: 0.25% Grok-4.20: 0.00% François Chollet just released ARC-AGI-3 -- the hardest AI test ever created. 135 novel game environments. No instructions. No rules. No goals given. Figure it out or fail. Untrained humans solved every single one. Every frontier AI model scored below 1%. Each environment was handcrafted by game designers. The AI gets dropped in and has to explore, discover what winning looks like, and adapt in real time. The scoring punishes brute force. If a human needs 10 actions and the AI needs 100, the AI doesn't get 10%. It gets 1%. You can't throw more compute at this. For context: ARC-AGI-1 is basically solved. Gemini scores 98% on it. ARC-AGI-2 went from 3% to 77% in under a year. Labs spent millions training on earlier versions. ARC-AGI-3 resets the entire scoreboard to near zero. The benchmark launched live at Y Combinator with a fireside between Chollet and Sam Altman. $2M in prizes on Kaggle. All winning solutions must be open-sourced. Scaling alone will not close this gap. We are nowhere near AGI. (Link in the comments)
Polski
@altryne That really felt like talking with something intelligent. Every single voice mode except for Sesame just feels like a TTS synthesizer, completely uninteresting to even listen to. And talking with them is just annoying.
English
This is what OpenAI has promised us with "advanced voice mode" 2 years ago!
We were robbed
English
@flowersslop Not a big loss. I trained a custom model on ~35 real songs I wrote back in the day, and what it generates often sounds like a chaotic mashup of random parts of my actual tracks switching every other beat, and can even stack multiple melodies in different keys at the same time.
English
KONSTYTUCJĘ III RZECZPOSPOLITEJ POLSKI STWORZYŁO AI? Taka przerażająca prawda wyłania się z przeskanowania jej tekstu za pomocą GPTzero. To oczywiście bzdura. Ale tyle są warte "skanery" tego czy coś jest napisane AI lub z pomocą AI. Inne wyjaśnienie: ktoś dysponuje maszyną do podróży w czasie.
Polski
@kimmonismus Anyone who read through model safety reports knows how dangerous these things could be. Thinking that super-intelligence will certainly be super moral, tame, would only cure cancer and we 100% could easily control it is just wishful thinking.
English
That's so stupid, they're seriously walking around with "Stop Skynet" signs.
They've watched too many doom and gloom sci-fi movies.
We still have a lot of educational work to do, friends.
Michaël Trazzi@MichaelTrazzi
On our way to OpenAI!
English
‼️OJAPIERDOLE‼️
Był update dosłownie 3 minuty temu w CS2 i zmienili że nie użyte ammo w magazynku przepada po reloadzie XD
Zmiana 10/10‼️
Polski
@thepolandnews_ Guys, Szon to z angielskiego Sean jak np. Sean Connery
Polski
BAJKA, PRINCESSA, A MOŻE SZON?
W Polsce rośnie moda na oryginalne imiona, a niektóre z nich nadawane są wyjątkowo rzadko.
Jak wynika z danych aplikacji BRAIAN, w 2025 roku najrzadziej nadawane imiona to m. in. dla dziewczynek: Apolina, Arwena, Daliana, Delfina, Dua, Emilija, Esmeralda, Henryka, Iza, Kajla, Ksymena, Mariola, Miłka, Oliwiia, Sława, Wanesa i Wiera, a dla chłopców: Aksel, Arek, Bronimir, Dobromił, Dyzma, Gaweł, Hipolit, Melchior, Palermo, Rajan, Riczard, Światosław, Tristan i Zbyszko.
Jeszcze rzadsze są imiona z rejestru PESEL. Wśród kobiet to np.: Bajka (2), Barbra (4), Chwalimira (6), Dąbrówka (193), Domasława (6), Dżanet (4), Dżoana (7), Halka (11), Helenka (10), Izaura (54), Jaga (41), Jaśmin (9), Jeżyna (6), Lidka (3), Nasturcja (12), Macieja (26), Princessa (17), Shakira (20). Wśród mężczyzn: Buba (23), Dżon (11), Edek (2), Kajtek (4), Lolek (9), Luidżi (2), Mirmił (2), Morfeusz (13), Mściwoj (10), Olimp (2), Ozzy (10), Sebastiian (12), Szaweł (3), Szon (4), Waldemard (15), Zawisza (41), Zefir (9), Zuber (2).
Jednocześnie z użycia zanikają imiona takie jak Wiesława, Józefa czy Dżesika, a do łask wracają Leon i Hanna.
Polski
@6GhosT9 Słyszałem, że designerzy z tych kilku AAA game studios uczestniczących w becie DLSS5 mówili, że są pod wrażeniem tego i wygląda to bliżej ich założeń jak chcieli by ich gry wyglądały bo zawsze musieli iść na kompromis i zaniżać grafikę by wszystko działało w realtime z dobrym FPS
Polski
Czy banieczka gamingowa zdaje sobie sprawę, że DLSS5 to jest narzędzie do wykorzystania przez developerów? A nie filtr AI nakładany na to co silnik wypluwa?
To trochę tak jakbyście narzekali na to, że Unreal ułatwia tworzenie postaci swoim Metahuman.
Polski
Important update: at the press conference I was able to talk to NVIDIA employees about DLSS5.
They assured me that DLSS5 makes no changes to the characters at all (!), no changes to the assets, and no "AI filter" is applied to the characters. It's solely the lighting that makes it appear that way; they are completely the same, unaltered assets.
Chubby♨️@kimmonismus
I am seriously so annoyed by all those AI haters. DLSS5 looks phenomenal; people hate on it cause „it’s AI“. That’s the sole reason.
English
@kimmonismus The sheer amount of ai generated comments here makes me think that most impacted jobs are social media accounts lol. No guys, when you convert to lowercase and remove em dashes it's still obvious you generated it with an LLM.
English
Andrej Karpathy has launched an analysis of AI on jobs. Karpathy is by no means interested in hype or exaggeration.
Using AI, he concluded that out of 143 million working people in the US, approximately 57 million are at high to very high risk of their jobs being negatively impacted by AI.
That's almost 40%. Let that sink in and consider what it means.
And then, in that context, consider that Meta is thinking about laying off 20% of its workforce for efficiency reasons.
Musk says universal high income is coming. I hope sooner rather than later.
Kaito | 海斗@_kaitodev
5 minutes ago, @karpathy just dropped karpathy/jobs! he scraped every job in the US economy (342 occupations from BLS), scored each one's AI exposure 0-10 using an LLM, and visualized it as a treemap. if your whole job happens on a screen you're cooked. average score across all jobs is 5.3/10. software devs: 8-9. roofers: 0-1. medical transcriptionists: 10/10 💀 karpathy.ai/jobs
English
@kimmonismus They want to start a new trend
English
"Very few people were opting in to end-to-end encrypted messaging in DMs, so we're removing this option from Instagram in the coming months. Anyone who wants to keep messaging with end-to-end encryption can easily do that on WhatsApp."
Let's see if WhatsApp will really remain end-to-end encrypted in the long term, or if this post will age poorly.
English
@DenisKalambury @thepolandnews_ Jako programista uważam, że ta liczba jest całkiem dobrze zaokrąglona
Polski
@thepolandnews_ Nikt by nie wysyłał 64 KG, dziwna liczba. Zamawia się okrągłe towary więc było albo 65 albo 70 :)
Polski
ALEŻ POMYŁKA
Pracownicy jednej z firm w powiecie brzeskim przeżyli duże zaskoczenie, gdy odebrali przesyłkę przeznaczoną dla swojego przedsiębiorstwa. Na palecie, zamiast zamówionych komponentów do produkcji urządzeń, znaleźli foliowe worki wypełnione marihuaną. Na miejsce wezwano policję z Opola i Brzegu, która zabezpieczyła prawie 64 kilogramy narkotyków.
Śledczy oszacowali, że czarnorynkowa wartość tej nietypowej przesyłki wynosi około 2,5 miliona złotych. Ze wstępnych ustaleń wynika, że narkotyki mogły pochodzić z transgranicznego przemytu i trafiły do firmy przez pomyłkę. W sprawie wszczęto śledztwo prowadzone przez brzeskich funkcjonariuszy pod nadzorem prokuratury.
Polski
@kimmonismus Where are these cures? Breakthroughs are close to none. Meanwhile, we're already seeing staff reductions. AI generates art - something that was supposed to be left for us to do after AI inevitably takes all jobs. AI is used in hacking for recon and to create attack tools quickly.
English
Bernie Sanders is advocating for an AI moratorium.
And I am completely against it, for the following reasons:
- Sanders focuses solely on the supposedly negative aspects, specifically the idea that AI could destroy all jobs.
- That's quite possible, but it's not an argument *against* AI, but rather an argument for realigning society. The fact that AI and robotics could take away tedious work is something I very much welcome.
- At the same time, the positive aspects are completely ignored: cures for (all) diseases, breakthroughs in energy research, longevity, and so much more.
- His concern is that AI could go "rogue." But first, he would have to argue why that's even possible; otherwise, it's pure resentment.
- Last but not least: it's impossible to implement a moratorium. China would seize this opportunity and try to overtake the US at all costs.
Sen. Bernie Sanders@SenSanders
We need a moratorium on AI data centers NOW. Here’s why.
English
@icreatelife AI Slop. I'd rather do it properly from the start, with vision, with intent, and know what the outcome is going to be, then craft it. No-one is impressed by this garbage, and it shows.
English
🎉 We just released Rotate Object in Photoshop (beta) 🎉
You can now rotate 2D images! 🤯
Then use Harmonize to add light and shadows, to blend it perfectly with the rest of the scene.
It's like Turntable in Illustrator, but instead of vectors, it's pixels in Photoshop!
English
English
No i po całej przewadze Google Antigravity. Pierwszy raz kiedy pomyślałem o anulowaniu subskrypcji Google Pro AI od @Google . @antigravity napluło użytkownikom w twarz. Jeszcze ten bezczelny marketingowy bełkot. XD
x.com/antigravity/st…
Google Antigravity@antigravity
We’re evolving Google AI plans to give you more control over how you build. Every subscription includes built-in AI credits, which can now be used for Antigravity, giving you a seamless path to scale. Google AI Pro is the home for the practical builder, hobbyists, students, and developers who live in the IDE and don't necessarily rely on an agent. This plan features generous limits for Gemini Flash, with a baseline quota included to "taste test" our most advanced premium models. Google AI Ultra serves as the daily driver for those shipping at the highest scale who need consistent, high-volume access to our most complex models. If you’re on Pro but need "extra juice" for a heavy sprint or deeper access to premium models, simply top up your AI credits to customize your plan. Keep building. Keep shipping.
Polski
@StonerchefPL Alkohol też tak alkoholicy tłumaczą, że to tylko na dobry sen. Nazywaj rzeczy po imieniu. To ćpanie. Dla mnie masz pełne prawo robić co chcesz ze sobą - nikt nie powinien ci łaskawie pozwalać czy zabraniać tego, co, kiedy i ile chcesz tego brać.
Polski
@errors_here O, tutaj mamy typ medycznego purysty, bo przecież nie można sobie gibona spalić na dobry sen bez pozwolenia od pośrednika.
Polski
💚 Czy zioło da się ĆPAĆ?
Tak, np. waląc wiadra albo opalając lufę po kilku paleniach.
To jest nic innego, jak ćpanie.
Nie ma nic wspólnego z kulturą konopną.
Prędzej z konopnym menelstwem.
Polski