L3t

Geopolityka/Giełda/AI/

[L3t @L3t94]

@FinansowyUmysl Tak to jest jak skupia się nie na biznesie dla którego dostarcza się rozwiązania, a na samym klepaniu kodu. Wkrętarka? Ja wolę śrubokręt!

[L3t @L3t94]

@vineerpasam accurate 💯

[Marek Łada @MarekLada_]

GPT-6 kręci całkiem dobre filmy, aktorzy zbędni. 😎

[L3t @L3t94]

@FinansowyUmysl Wystarczy zostać mentorem AI xd

[L3t @L3t94]

@mickpll Post w stylu nie wiem to się wypowiem?

[L3t @L3t94]

@Tech2Wild No vision, no mtp. I will stay with Ornith1.5

[L3t @L3t94]

@alexocheema Low RAM/VRAM. I guess any local amout is low.

[L3t @L3t94]

@mfranz_on Check ornith1.5. For some task may be better.

[L3t @L3t94]

@superalesha @QwenDevs +1

[L3t @L3t94]

@FinansowyUmysl UOP to umowa śmieciowa. Na b2b ludzie chociaz maja świadomość ile hajsu idzie w podatki. Plus mają jakikolwiek wybor co do formy opodatkowania. UOP to taki blue pill z matrixa.

[L3t @L3t94]

@yacineMTB Robots...

[L3t @L3t94]

@clauslimenkov @realtek12345 @OpenAI @AnthropicAI trafne :)

[L3t @L3t94]

@FinansowyUmysl Ale creepy

[L3t @L3t94]

@FinansowyUmysl Różni ludzie mają różne doświadczenia z AI. Dlaczego? Bo w jednym repo jest duża jakość i AI też generuje jakościowy kod. A w drugim jest koszmar i AI potęguje istniejący syf :)
Najważniejsze to rozumieć, że to na ten moment jest narzędzie. My decydujemy jak zostanie wykorzystane

[Alex Finn @AlexFinn]

Absolutely insane. Major breakthrough today

Moderna used AI to design an mRNA cancer vaccine that succeeded in a phase 3 trial

It's not approved yet and it's just a treatment, but it's a step towards AI eventually curing cancer

It's incredibly sad people are trying to stop this innovation by (falsely) claiming AI data centers are drinking all our water

That AI is taking all the jobs

That AI is just a way for billionaires to get rich

All while the technology gets us closer and closer to cures for all diseases

I don't know which foreign country has convinced half of America that AI is the devil, but hopefully this "AI is drinking the water" trend ends sometime soon

We are on the cusp of greatness

[Ornith @ornith_]

Aloha! 🌺Introducing Ornith-1.5, a family of open-source LLMs spanning 9B Dense, 35B MoE, and 397B MoE, trained with self-improving strategies.

It achieves state-of-the-art performance among open-source models of comparable size and delivers performance comparable to Claude Opus 4.8 across reasoning, agentic, and coding tasks:
✅Terminal-Bench 2.1 (86.1)
✅SWE-Bench (86 on verified, 65.1 on pro, 79.6 on Multilingual)
✅DeepSWE (56)
✅HLE (44.6)
✅ClawEval (81.4)
✅Tool Decathlon (71.2)

Ornith-1.5 takes a major step toward training foundation models through end-to-end self-improvement, extending the self-scaffolding strategies introduced in Ornith-1.0 into a more complete self-improvement loop: the model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve.

All models, along with their quantized versions (FP8, GGUF, MLX, and NVFP4), have been released under the MIT License, enabling unrestricted commercial and research use.
📘Tech Blog: https://t.co/OZ63scRWLB
🤗Huggingface: https://t.co/mGJLwhrQOM

[aipulsedaily @aipulseda1ly]

Opus 5 has completely destroyed my expectations for what an LLM can do.

26 billion tokens. 4 weeks of feeding it Claude accounts nonstop. 124 seconds of 4K, 2978 frames, one camera, zero cuts.

Nothing imported. Nothing sculpted by hand. Zero texture files. Every surface is code it wrote itself.

17,707,774,735 triangles evaluated every single frame. 52.7 trillion across the film.

It rented 3 RTX 5090s on its own through middleware another Claude agent built. 231 GPU hours for the final render.

The render cost $229. The thinking cost $20,740. The 4K film was the cheap part by 90x.

It also rendered the car inside out for weeks and none of us caught it. A back facing solid fills the same silhouette, so it still looked like a car.

Audio is the one thing it never got. Rebuilt five to seven times and I reverted to the original every time.

A year ago none of this was possible.

[AJ @ItsmeAjayKV]

Fk, they remove it, looks like we have to wait a bit more and see if we will be getting 3.8-35b

Never say never, so still hopeful.

[AJ @ItsmeAjayKV]

See that everyone

We will also be getting a Qwen3.8-35B

LFG @QwenDevs 👏🏻

[Hikari∣LocalLLM⚡ @Hikari_07_jp]

IQ4 is superior to NVFP4!!
It seems to achieve high quality through explicit bit allocation on a tensor-by-tensor basis. They have also released similar quantization for Ling 3.0 Flash and Glimmer 30B.

Following them is recommended.