The only programmers who are really in trouble with the competition from artificial intelligence are the ones who insist the world today isn't all that different from the one we lived in last year.
Embrace the paradigm shift or perish.
(This has always been true.)
HexLab Software retweeted
Mentre il TG1 ci informa, con tanto di servizio, che il calciatore Haaland si è tagliato i capelli in Cina prosegue il mega evento dedicato alla robotica: due record umani sono stati battuti da robot umanoidi. Fino a pochi anni fa questi robot faticavano persino a restare in piedi e costavano circa 120 mila dollari. Oggi corrono più veloci di un atleta umano e costano quasi dieci volte meno. Se rappresentassimo tutto questo su un grafico, vedremmo una crescita esponenziale delle prestazioni e al tempo stesso un crollo altrettanto rapido dei costi. L’apprendimento dei robot corre molto più veloce di quello che possiamo immaginare
When I say Omarchy runs great on a potato, I really mean it! Here is an i5-2520M ThinkPad X220 with 2GB RAM from 2011(!!!) running Omarchy smooth. It uses just 890MB at idle, and just 1.3GB with several Chrome tabs, YouTube 1080p, AND OBS running 🤯
HexLab Software retweeted
Grok 4.6 on extra high thinking mode now achieves #1 score on CursorBench!
Grok 4.6 just took the #1 spot on CursorBench 3.2.....and the efficiency is insane
Here's the cost comparison:
• Grok 4.6 Extra High — 70.8% | $2.81/task
• Fable 5 Max — 70.5% | $17.32/task
• Opus 5 Max — 70.0% | $8.23/task
• GPT-5.6 Sol Max — 67.2% | $5.69/task
Grok achieved the highest score while costing roughly 6X less than Fable 5 Max and nearly 3X less than Opus 5 Max per task
That’s what makes Grok so powerful for agents
Top-tier intelligence is great.....but top-tier intelligence that can keep working across long coding tasks without burning ridiculous amounts of compute is even better
Grok’s agentic coding efficiency is insane
I have never tried Omarchy so far but there are a few reasons why I think what DHH is trying to do with Linux on the Desktop could be significant. There is the right mix of AI help with many boring and costly tasks (drivers for instance), MacOS dropping in quality, and the fact the only reason why Linux never succeeded in the Desktop is that UX requires an opinionated development path from some entity able to also fuel it, not the distributions setup Linux historically had. For all this reasons, a few starts aligned, so this could be the right time. Also ROCm / Vulkan / CUDA GPU hardware options are evolving, and could end creating a compelling alternative to Apple hardware for local inference: but this is also the major risk, if Apple wins the local inference hardware battle, MacOS could become a forced choice for many.
The Omarchy manual is now searchable thanks to @jorgemanru. Super fast too! Just hit /, do you search, hit / again for another. omarchy.org/manual/
DeepSeek v4 PRO running in DwarfStar on a DGX Station. Can go faster, already peaks 50 t/s without DSpark: dynamic VRAM/RAM allocation of experts based on past statistics, 4k t/s prefill on big chunks. VRAM/RAM layer-major streaming. Mixed RAM/VRAM inference requires care.
HexLab Software retweeted

Tibo@thsottiauxAug 12
United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
I previously promised a reset for every 1M in additional active users for Codex, until 10M.
We blew past that and have been silent since 10M. Little surprise for you tomorrow.
HexLab Software retweeted
👋🏼ADIÓS A LOS ANIMADORES DE JUEGOS
Ya pueden ir aplicando a MCDONALD.
NVIDIA acaba de cargarse 30 años de tecnología de animación de golpe.
Sin blending.
Sin foot-locking.
Sin detección de colisiones.
Sin UNA SOLA transición hecha a mano.
Y se ve mejor que lo que hace la industria
Se llama MotionBricks.
Un solo modelo neuronal, 350.000 habilidades de movimiento, 15.000 FPS, 2 milisegundos de latencia.
Se usa como quien apila bloques:
→ Le dices velocidad, dirección y estilo, y compone la locomoción sola
→ Herido, zombi, saltitos, lateral. Zero-shot, sin reentrenar nada
→ Cambias de estilo en caliente, en mitad del movimiento
→ Para agarrar un objeto pones cuatro keyframes aproximados y la red rellena cómo se acerca, cómo lo coge y cómo termina
→ Cada vez sale ligeramente distinto, como una persona real
Sin tocar un grafo de animación. Sin saber animar.
Y lo más fuerte: el mismo modelo ya mueve robots humanoides reales dentro del stack GR00T de NVIDIA.
Personajes de videojuego y humanoides de carne y metal, con la misma red.
Enlace en el primer comentario 👇
HexLab Software retweeted
Vibed out an app today and I haven't looked at the internals at all
Things are going wrong with it, I'm scared and frustrated
Then I remembered /improve-codebase-architecture
What a lovely, warm bath of a skill
HexLab Software retweeted
New in Claude Code: your sessions can now message each other.
Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.
HexLab Software retweeted
Introducing Claude Opus 5.
It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.
Replying to @kai_fell
I worked at Redis till 2020, so the irrelevancy phase was just from 2020 to 2024 (when I rejoined). Also I'm older now. Please could you fix, without the help of AI, your graph to reflect the current time line? Also you may want enjoy my Vector Sets implementation. It is hand-written, a few months ago, and is one of the best HNSW implementations out there.
HexLab Software retweeted
JUST IN: A new Chinese AI model from Zhipu AI reportedly matches Claude Mythos’ performance at finding security bugs.
Replying to @derekcollison @ziglang
Could I (respectfully!) question that? I believe that today people that, in the long term, will not embrace AI, will create projects that will not survive long term. I'm all for writing code by hand for artistic reasons (think at poetry), but as models get more powerful, the no AI policy will condemn every large project to irrelevance soon or later.
This does not mean to lack respect of other people's view. AI code right now is sometimes a downgrade compared to the best coders. But we already crossed the point in a definitive way where the advantage definitely overcome the disadvantages.
HexLab Software retweeted
My second short story release in English is ready: Tales of Illustrious Computer Scientists: Iola Varga, nun and computer scientist. invece.org/iola.html
Prediction: now that with GLM / DeepSeek & more we will see very good LLMs available open weight, and even more when the GPU/RAM insanity will be contained, the open source agents landscape will create *much* better tools that we are used to use today. For instance:

















