Piotr Prekurat

sorry, can't talk. new model dropped

[Piotr Prekurat @sensei_prek]

@OpenAI

[Piotr Prekurat @sensei_prek]

🚨 google just shipped gemini omni 1.2 flash for video gen

- in ai studio and the gemini api
- 30% cheaper than omni flash 1.1
- video up to 15s, multi-turn edits in chat
- more camera control, less visual drift

omni flash just got a quiet mid-cycle upgrade.

[Google AI Studio @GoogleAIStudio]

introducing gemini omni 1.2 flash, now available in ai studio and the gemini api

what's new:

- fast & 30% lower cost
- longer clips
- edit in conversation
- more natural speech
- more controllable shots

[Piotr Prekurat @sensei_prek]

🚨 openai just launched gpt-6 cyber for security research

- 948 linux kernel lpe patches, 81 novel firefox bugs
- exploitgym 48.5% vs 32.1% on 5.6 cyber
- already at mozilla, palo alto, cisco
- daybreak red first, expanding later

cyber dropped before 6.2 astra and sol.

[OpenAI @OpenAI]

Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.

The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.

GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.

https://t.co/Nzab3Z1Td6

[Piotr Prekurat @sensei_prek]

@sufrocrata @OpenAI Lmao 💀

[Piotr Prekurat @sensei_prek]

@valigo 100p agreed. People wont automate shopping online, they just enjoy the process of scrolling through products, comparing options etc. Actual buying was never the goal

[Piotr Prekurat @sensei_prek]

I feel like we're reaching a critical mass where everyone launches things. None of it will matter in 1-2 years when big labs commoditize enrire tech space.

[Piotr Prekurat @sensei_prek]

🚨 openai just shipped images 2.6 with animated gifs in chatgpt

- make a gif from a prompt or uploaded photo
- edit frames and speed in chat
- api: gpt-image-2.6 flare (as fast as 6s) and sunburst
- image gen ~50% faster, per openai

[OpenAI @OpenAI]

Images, upgraded—introducing ChatGPT Images 2.6.

Now with GIFs and reference images.

Plus, two new models are coming to the API with 50% faster image generation.

[Piotr Prekurat @sensei_prek]

🚨 5 new gemini checkpoints codenamed "hydrogen" reportedly spotted on lmarena

- flash-class model, per leaker @chetaslua
- claimed speed: 300-500 tok/s
- no confirmation from google yet

if those speeds hold, flash stops being the budget pick and starts being the default.

[Piotr Prekurat @sensei_prek]

@Kimi_Moonshot downloading before anyone changes their mind

[Piotr Prekurat @sensei_prek]

🚨 moonshot just released the open weights of kimi k3.1 pro

- official announcement, no benchmarks or pricing yet
- the information previously reported k3.1 pro was distilled from gpt-6 astra outputs

if that report is right, openai's flagship just became the teacher for an open model. does that change whether you'd run it?

[Piotr Prekurat @sensei_prek]

@woocash_j Agree. I’m 23 and see it firsthand: growing up with post-communist parents, a “safe job” was the definition of success. The bigger issue is scale: most founders here think local, not global, and the ones who think big often leave. Long way to go, but it’s changing.

[Piotr Prekurat @sensei_prek]

@GeorgeJeffersn my first hustle was building server spawns for money. was like 11 yo, first internet money. things never been the same since

[MTS @MTSlive]

SITUATION BREWING: Sam Altman says that the next GPT, the internal model that solved Navier-Stokes, will be released "soon."

[Rikiya Takehi @rikiyatakehi]

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.

We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq

[Piotr Prekurat @sensei_prek]

🚨 Multiple Grok Bot users report higher usage and hitting limits early.

Expect less spam on your timeline ‼️

[Piotr Prekurat @sensei_prek]

🚨 anthropic's official mythos page now reads "claude claude mythos 5.5"

the rest of the page still describes mythos 5.1. no announcement yet, but the header changed first. a launch looks close.

[Piotr Prekurat @sensei_prek]

@kimmonismus claude claude mythos 5.5. so good they named it twice

[Piotr Prekurat @sensei_prek]

@LuminaBench astra in a whale costume at 1/10 the price. i'm in

[Piotr Prekurat @sensei_prek]

deepseek v4.1 pro is already showing up as the target model in searchqa benchmark configs

flash was the appetizer. no date yet, but november is starting to look crowded

[Lumina @LuminaBench]

DeepSeek V4.1 Pro is showing up in benchmark setups

This is V4.1 Pro as the target model on the SearchQA Benchmark

I'm guessing we've had to wait this long for the Pro version so they could distill GPT-6 Astra just like they did with Flash, but i'm certainly not complaining