Thomas Tränkler

2K posts

Thomas Tränkler banner
Thomas Tränkler

Thomas Tränkler

@ttraenkler

Cognitive Neuroscientist, AI and Fullstack Software Engineer, Founder of @loopdive

München, Bayern Katılım Haziran 2009
3.1K Takip Edilen529 Takipçiler
Sabitlenmiş Tweet
Thomas Tränkler
Thomas Tränkler@ttraenkler·
JS² compiles JavaScript to WebAssembly ahead-of-time, sandboxing untrusted modules in your supply chain – running in JS engines or standalone – no interpreter or engine bundled. Goal: 100% ECMAScript & NPM compatibility; Method: spec/test-driven agile team of agents against spec + 100.000 unit tests, passing 75.4% of test262 in a JS host, 43.8% standalone. Module sandboxing limits blast radius for supply chain attacks or AI generated code Lightweight with no bundled interpreter or engine, starting at 100 bytes Fast, typically being an order of magnitude or more faster than interpreters or engines without JIT embedded in the module Garbage Collection natively supported by Wasm 3.0, builtin and robust instead of avoiding or embedding Help us bringing this vision of a new substrate for running JavaScript as lightweight, composable modules to life – by reviewing, testing, contributing, or donating! js2.loopdive.com @loopdive #JavaScript #TypeScript #WebAssembly #sandboxing #security
Thomas Tränkler tweet media
English
4
6
15
2K
Thomas Tränkler retweetledi
👩‍💻 Paige Bailey
👩‍💻 Paige Bailey@DynamicWebPaige·
You can now fine-tune models on your personal laptops, and run the latest @GoogleGemma 4 models with as little as 3GB VRAM (🤯). 👇Stellar work, as always, from the @UnslothAI team!
Unsloth AI@UnslothAI

Introducing Unsloth for AMD 🚀 You can now train & run LLMs on your AMD hardware • We collaborated with AMD to enable you to train & run 500+ models on AMD GPUs • Works on Windows, WSL, Linux • Train Qwen, Gemma on 3GB VRAM GitHub: github.com/unslothai/unsl… Works on Radeon, Instinct, Ryzen and data center GPUs with up to 2× faster with 70% less VRAM and no accuracy loss via our custom Triton kernels and math algorithms. We also support optimized ROCm builds for GGUF & Safetensors inference. Unsloth is an open-source local UI for faster LLM training and inference, with tool-call healing, code execution, secure web search, remote APIs, and HTTPS deployment. Connect local models to Claude Code, Codex agents and run the latest Kimi, GLM, DeepSeek, Qwen3.6, and Gemma 4 models. 🔗Blog + Guide: unsloth.ai/docs/basics/amd

English
5
8
67
7.2K
Thomas Tränkler retweetledi
Qwen
Qwen@Alibaba_Qwen·
Qwen3.8 is launching and going open-weight soon!🌐 With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5. You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out. Can't wait to hear what you build. Stay tuned! 🚀  Token Plan international:qwencloud.com/pricing/token-… China:platform.qianwenai.com/pricing/token-…
Qwen tweet media
English
1.3K
3.3K
24K
7.8M
Thomas Tränkler retweetledi
Ryan Lopopolo
Ryan Lopopolo@_lopopolo·
I distilled myself and my work to twelve theses of harness engineering. This represents the last year of my work. This is the practice and technique. Go nuts fam. github.com/lopopolo/harne…
Ryan Lopopolo tweet media
English
21
31
602
73.4K
Thomas Tränkler retweetledi
Vercel
Vercel@vercel·
this is who runs this account
English
31
25
719
62.1K
Thomas Tränkler retweetledi
Claude
Claude@claudeai·
Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit. Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.
English
5.3K
6.3K
53.5K
27.6M
Thomas Tränkler retweetledi
Peter Steinberger 🦞
Are we still talking loops or did we shift to graphs yet?
English
1.2K
380
7.6K
2.8M
Thomas Tränkler retweetledi
Kimi.ai
Kimi.ai@Kimi_Moonshot·
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
Kimi.ai tweet mediaKimi.ai tweet media
English
1.7K
7.6K
56.8K
23.2M
Thomas Tränkler retweetledi
Kimi.ai
Kimi.ai@Kimi_Moonshot·
Meet Kimi K3
Türkçe
748
2.7K
30.6K
10.1M
Thomas Tränkler retweetledi
Thinking Machines
Thinking Machines@thinkymachines·
Today, we are introducing Inkling. Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available. thinkingmachines.ai/news/introduci… Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
English
561
2K
14.8K
7.9M
Thomas Tränkler retweetledi
Gustavo Noronha
Gustavo Noronha@gustavokov·
Linus jogou a real: AI é útil e vai ser usada no desenvolvimento do kernel Linux, quem achar ruim pode fazer fork e vazar
Gustavo Noronha tweet mediaGustavo Noronha tweet media
Português
134
741
7.4K
762.9K
Thomas Tränkler retweetledi
INK
INK@0xInk_·
To any game developers seeing this post: all I want is a solid exploration game with a grappling hook mechanic (Last Oasis style) someone take the concept in this video and make it a reality. I am ready to throw my money at this!
English
300
843
9.9K
543.8K
Thomas Tränkler retweetledi
Ben Davis
Ben Davis@davis7·
Been going back to Pi, forgot how absurdly good it is...
Ben Davis tweet media
English
66
33
1.1K
75K
Thomas Tränkler retweetledi
Claude
Claude@claudeai·
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English
7K
7.3K
74.1K
27.7M
Thomas Tränkler retweetledi
Chris Tate
Chris Tate@ctatedev·
Introducing Native SDK The toolkit for building native apps → Hot reload → Markup + Zig → Instant launch → macOS, Windows and Linux → GPU engine built from scratch → Built-in design system and themes → Custom components + design tokens
Chris Tate tweet media
English
143
160
2.5K
577.2K
Thomas Tränkler retweetledi
Brayden
Brayden@BraydenWilmoth·
Introducing Cloudflare Drop Drop your folder in the browser and deploy it instantly on Cloudflare. Your website... milliseconds away from users on region: earth No account needed. Deployment is active for 60 minutes, then expires unless you claim it. cloudflare.com/drop/
Brayden tweet mediaBrayden tweet media
English
288
845
9.3K
1.4M
Thomas Tränkler retweetledi
Solomon Hykes
Solomon Hykes@solomonstre·
If your AI writes the word "honest", immediately kill the session and throw away whatever it produced.
English
22
12
137
18.1K
Thomas Tränkler retweetledi
OpenAI
OpenAI@OpenAI·
Introducing GPT-Live, a new generation of voice models for natural human-AI interaction. Rolling out in ChatGPT starting today. You’ll want to turn the sound on for this one.
English
1.3K
2.2K
24.5K
7.9M