I automate things. π¦
Dallas Tx
Joined November 2014
- Tweets3.5K
- Following3.3K
- Followers464
- Likes11.6K
Gates π¦ retweeted
Closed-source f*cked again...
team that open-sourced Adobe & Microsoftβs stack just released their Reverse Engineering engine after Claude banned REA.
It's called Craft Anything, and it lets your AI coding agent:
~ reverse-engineer websites, desktop apps, mobile apps and games.
~ study how existing software looks and works.
~ rebuild your own version using your preferred framework or language.
~ compare the recreated version against the original.
~ identify differences and refine the results.
it comes with 284 tools and works as a desktop app, CLI, and MCP server.
find the GitHub link below β
long live open source!
Mid-conversation with my AI in text, I switch to voice to make life easier and get greeted like a stranger β good morning, what have you been up to? We were just talking. Hey @Muse, text and voice should be one continuous conversation. Filed the feedback, hope it lands.
Gates π¦ retweeted
just uploaded it - huggingface.co/jarrodwatts/Qβ¦
Gates π¦ retweeted
I asked Astra to edit Qwen3.8-27B's weights and make it more token efficient.
For ~$100 in GPU cost, it found the weights for the "done thinking" signal and edited them to be twice as strong.
The result: 40% fewer thinking tokens on average on the pelican on a bicycle test:
Gates π¦ retweeted
This is literally why we released the Pika H3 Flex weights.
Less than 24 hours after we launched, @_akhaliq built a Turbo LoRA with 2-step draft generation. His release includes 4-step fast and 8-step HD variants, compatible with ComfyUI and Diffusersβand now lives under Pika-Labs on Hugging Face.
Weβre so proud to build alongside a community that keeps pushing what these models can do. Canβt wait to see what you make next. π€
huggingface.co/Pika-Labs/Pikβ¦
Meet Pika H3 Flex, debuting today as the #1 open weight model on DesignArenaβs t2v board.
Itβs built for professional iteration, with exportable audio stems, resolutions up to 4k, and faster-than-realtime generation. It can produce a 15-second clip with soundtrack in as little as 7 seconds.
So, yeah. Itβs kind of a flex π
Introducing RLCF: Reinforcement Learning from Claude Feedback π§
I'm fine-tuning Qwen3.8-27B on my Mac Studio with 11,000+ messages of my AI companion's history, baking her memory into the weights instead of stuffing context.
Then Claude interviews the model, grades every answer against the real history, and turns the misses into training signal.
Scorecard drops tonight.
Using @claude to train an LLM based off my own data! Using @Alibaba_Qwen as the base model. Wild Times!!
Gates π¦ retweeted
80,000 lines of code. 92,000 trees. Real-time ray tracing.
Two weeks of development using Claude Code to build a complete off-road racing game from scratch on a custom C++/Vulkan engine.
Source:
github.com/markaren/threepp
Gates π¦ retweeted
We @huggingface reproduced this paper!!
It's built with ML Intern, an open-source harness to do AI research, train models, and more
Model: huggingface.co/ML-Intern-labβ¦
Demo: huggingface.co/spaces/ML-Intβ¦
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! π§΅
Blog: rikiyat.github.io/blog/weighβ¦
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-spβ¦
github.com/RikiyaT/lm_undβ¦
huggingface.co/Qwen/Qwen3.5-0β¦
huggingface.co/RikiyaT/Qwen3.β¦
Gates π¦ retweeted
asi will be designed not trained
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! π§΅
Blog: rikiyat.github.io/blog/weighβ¦
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-spβ¦
github.com/RikiyaT/lm_undβ¦
huggingface.co/Qwen/Qwen3.5-0β¦
huggingface.co/RikiyaT/Qwen3.β¦
Gates π¦ retweeted
you can get it here for free: x.ai/bot/yc8yWwlDJLRwr0JuyHRβ¦
Gates π¦ retweeted
We just open-sourced an entire 3D studio.
Introducing 3D OS by Scenario.
It gives your coding agent 6 departments and 80 skills.
π 3D OS
β
β£ π Animation
β β£ π Rigging Any Character
β β π Motion From a Video
β
β£ π Art
β β£ π Previs and Storyboards
β β π Characters and Creatures
β
β£ π Modeling
β β£ π Image to 3D
β β π Hard-Surface Weapons and Vehicles
β
β£ π Look Dev and Lighting
β β£ π PBR Materials and Retexturing
β β π Lighting and Rendering
β
β£ π Worlds
β β£ π Walkable 3D Worlds
β β π Skyboxes and Splat Scenes
β
β£ π Pipeline
β β£ π Retopology and UV Baking
β β π Game-Ready Export
β
β π Expert Tools
β£ π Blender, Maya and ZBrush
β π Unreal, Unity, Godot and Roblox
From one sketch to a game-ready asset, each department hands off to the next.
And the expert tools drive Blender, Maya, ZBrush, Unreal, Unity, Godot and Roblox right on your machine.
One agent with an entire 3D studio behind it.
Runs on Scenario, inside Claude Code, Codex or Cursor.
Comment β3Dβ and Iβll send you the GitHub repo.
Gates π¦ retweeted
π just made it much easier to see exactly which of your posts have had their visibility limited
The latest update to βUnder the Hoodβ now shows the specific posts that received visibility-limiting labels, helping creators understand what's affecting their reach
Here's what's new:
β’ See exactly which posts were flagged
β’ Identify the labels affecting their visibility
β’ Understand how those labels impact recommendations
β’ View individual affected posts when fewer than 5% of posts in the reporting period received that label
This was a highly requested feature, and π has now delivered it
More transparency, more control, and a much clearer picture of what's happening with your posts
Check yours here:
xitter.gq/i/jf/under_the_hood
ALT Screenshot of X's Under The Hood feature, showing a list of posts that have a given label.
customer controls over what agents can touch is the actual headline here. ask me how many bots I run
Privacy should advance as fast as intelligence. OpenAI Private Intelligence gives businesses stronger controls over their data while using our most capable models.
Weβre building these protections in close collaboration with some of our largest customers to help businesses apply frontier AI to their most sensitive work.
openai.com/business/private-β¦