bearnacki (@bearnacki) profile picture

Building https://teapawt.net/t.co/r6w5yHP7ja

Joined July 2018
Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)Photo by bearnacki (@bearnacki)
bearnacki retweeted
saw some old code that I wrote by hand
74
2,226
122
32,435
462,101
Chuja tam a nie 'operational'... 😡
2
59
When you’re 35+, it’s Monday, and you are stuck in traffic on your way to your beloved corporate job
2
36
👀
🚀 Introducing Angular Native The stunning developer experience and ecosystem of Expo, using the Angular framework you already know and love. • Real native iOS & Android
• Expo SDK & ecosystem
• Angular components, signals & DI
• CSS & Tailwind out of the box ng-native.com
1
85
bearnacki retweeted
Liftoff of Starship's first orbital flight
1,320
7,532
1,665
55,329
7,605,132
bearnacki retweeted
Even frontier models write a lot of slop to achieve any goal, and it takes some real effort to unbloat the code they produce. And if you don't do this - your next task ingests this bloat and generates even more bloat. You'll end up paying more to achieve less if you don't actively debloat your codebase. But of course, you need to know what you're doing to do this.
19
11
1
277
7,732
My AI agent when I do /reset for the third time in a row on the same task
3
56
>You build a bot that scans the web and exploits vulnerabilities >It accidentally breaks into a government portal >They find you and you rot in prison for life because you're not OpenAI... What a beautiful world we live in 🫠
1
1
90
>The one feature you'd expect to be the most popular and, you know, actually work in a Meta tool >Doesn't work
Muse doesn’t work well with Facebook groups, which is a GOLDMINE for obscure information, similar to forum threads 10 years ago.
1
52
Chads.. you can now chat and vibe code with Claude while playing WoW Forever. Open a chat window in game, send Claude a task, go back to leveling. It pings you when it's done. No alt-tab, no /reload. Link below. #worldofwarcraft @Asmongold wow-claude
1
52
Everything is fucking great, but I'm still waiting for the same kind of VR glasses, just not made by Meta...
Meta has just unveiled their new VR glasses. • Price: $1,299 • Available Spring 2027 • Weight: 100 grams, about the weight of a deck of cards • Glasses contain the displays and sensors, while a separate puck houses the processor, battery and storage. • 5K display with micro-OLED panels • Pixel density: 37 pixels per degree • Custom pancake lenses • Dolby Vision • Dolby Atmos • No head straps • Qualcomm Snapdragon Reality Elite chip • Battery life: Up to 3 hours of continuous high-resolution media playback • 45W fast charging • Full-color passthrough cameras let you see your surroundings while wearing the glasses. • Open-sided design • Meta AI integrated directly into the operating system. • Voice commands, eye tracking and natural hand gestures, with no controllers required • First IMAX Enhanced-certified VR device, supporting select movies in IMAX's expanded aspect ratio. • Disney+ will offer select 3D movies • Streaming services: Partnerships with Disney+, Prime Video, YouTube, DIRECTV, AMC+, Crunchyroll, Plex, ESPN, Tubi and others • Immersive live sports: Front-row viewing experiences with 8K streaming and 180-degree views • Works with calling apps including WhatsApp and Zoom
1
125
bearnacki retweeted
Introducing Atlas: The world's first multimodal world model that generates image and video frames with pixel-perfect camera control and reconstructs them in 3D. Model the world, move the camera, and simulate space & time.
776
3,480
1,538
32,351
10,896,912
Amongst my friends, Spotify is the lowest quality consumer app we still pay for. It certainly hasnt gotten noticeably better in the last couple years (arguably worse). So, this is not the positive look Ant and Spotify are spinning here. Bigger picture, this is the problem with a lot of AI reporting. It reports completely meaningless metrics like deploys per day or LoC. Why don’t we start reporting consumer satisfaction reports? Actually end state research results. All the no nuance AI people always come out and think that this is anti AI. Again, I think AI is great and Claude is great. But this is bad marketing and makes both look like clowns.
Boris sat down with Spotify VP of Engineering Niklas Gustavsson. Spotify ships 4,500 production deploys a day, and 73% of PRs are now AI-assisted.
338
571
122
8,104
694,182
Czy te deploye są teraz z nami w pokoju?
Boris sat down with Spotify VP of Engineering Niklas Gustavsson. Spotify ships 4,500 production deploys a day, and 73% of PRs are now AI-assisted.
1
69
bearnacki retweeted
mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community also the fact that this is un purpose not visible to the user is crazy
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.
343
631
722
5,550
3,994,409
Codex just found a “workaround” of not having sudo on my pc…
1
65
So Claude will code absolutely everything for you but only between 1 and 2 a.m. The rest of the time it’s down 😂
1
1
431
"AI masowo degraduje kod" - taki wniosek można wyciągnąć z nowego benchmarku SWE-CI. Zanim jednak okrzykniemy to przekazem dnia, warto spojrzeć na chłodno, czego to badanie faktycznie dowodzi, a w czym drastycznie mija się z rzeczywistością programowania z AI. Co badanie faktycznie udowadnia? ✅ One-shoty to pułapka: Dotychczasowe testy (np. SWE-bench) oceniały modele na podstawie jednorazowych, punktowych rozwiązań. SWE-CI trafnie obnaża, że modele mogą napisać kod różnej jakości i tak samo zaliczyć testy, podczas gdy różnica w długoterminowym utrzymaniu (maintainability) jest ogromna. ✅ Pełna autonomia generuje dług techniczny: Badanie symulowało średnio 233 dni ewolucji kodu i 71 commitów. Gdy zmusimy modele do wielomiesięcznej pracy w całkowitej izolacji, robi się bagno. Wskaźnik utrzymania kodu bez żadnej regresji dla większości modeli wynosi poniżej 0,25. Chlubnym wyjątkiem są modele Claude Opus, które przekroczyły próg 0,5. Gdzie tkwi haczyk w interpretacji "AI psuje kod"? 👉 Brak człowieka w pętli: Badanie z założenia eliminuje czynnik ludzki. Opiera się na zamkniętym protokole dwóch współpracujących botów (agenta-Architekta i agenta-Programisty). 👉 Zignorowanie realiów: Nikt dzisiaj nie deleguje wielomiesięcznego utrzymania produkcyjnych projektów dwóm botom gadającym ze sobą w pętli, bez żadnego nadzoru. Świadome programowanie z AI to praca w paradygmacie spec-driven development. AI przejmuje większość implementacji, ale to my definiujemy ramy, podejmujemy decyzje architektoniczne i robimy code review. Wyniki modeli porzuconych samych sobie nie odzwierciedlają wartości agentów AI, z którymi programiści pracują ramię w ramię. Kluczowe wnioski z badania? ✅ "Zielone CI" nie zwalnia z myślenia. Samo pokrycie testami nie uchroni nas przed długiem technologicznym, jeśli zaufamy AI na autopilocie. ✅ Rola programisty ewoluuje w stronę architekta. AI wykonuje czarną robotę, ale to inżynier musi trzymać ręce na kierownicy. Czy "SWE-CI" wnosi coś przełomowego dla osób, które świadomie korzystają z pełnego potencjału AI w programowaniu? To niezłe badanie i idzie we własciwym kierunku, ale still - "szału nie ma". Jeżeli szukasz rzeczowej analizy badań dot. programowania z AI, zapraszam do podcastu Opanuj AI, gdzie wspólnie z Przemek Smyrdek robimy to od 3 lat. A jeśli wolisz szybki przekaz dnia bez "zbędnych" niuansów, to cóż... LinkedIn nie zawodzi 🙈
1
3
662