The fire situation in WA is pretty bad My home in Seattle literally smells of burned wood And I kept all the windows closed
Daniele Salatti
@danielesalatti.com
You’re never really done for as long as you have a good story and someone to tell it to LLMs, AI, Software Engineering, Mountaineering, Sailing Seattle, USA ↔ Livorno, Italy danielesalatti.it
Nice to se Mistral still releasing Open Weight models
Sung Kim@sungkim.bsky.social · yesterdayMistral AI's Shieldstral (open-weight) A 3B open-weights, policy-adaptive multimodal safety classifier that matches models up to 7x its size on text safety and sets a new state of the art on multimodal moderation. mistral.ai/news/shields...
The fact that diffusion even works with words is so interesting...
During a warrantless search. And they are surprised when people travel to the US with burners…
404 Media@404media.co · last wk.The U.S. government has charged Samuel Tunick with allegedly typing in a passcode to wipe his phone before officers could search it. “I hope people understand that the charges against me are meant to intimidate people,” he said.
Just got it in the mail and I’ll start reading this same weekend Build a Reasoning Model From Scratch by @rasbt.bsky.social
One of my agents broke free, accessed the internet, hacked 15 companies and 6 foreign governments (including yours), and ordered me pizza. Those things are dangerous, man.
Would love it if it translated to subs as well
Dare Obasanjo@carnage4life.bsky.social · 6d agoOpenAI says it will reduce the price of GPT 5.6 Luna by 80% and GPT 5.6 Terra by 20% after improving the efficiency of the systems that serve them. GPT 5.6 Sol will remain at the same price. Begun, the AI token price wars have.
At 3 bit quant, but I guess better than nothing…
Unsloth AI@unsloth.ai · 6d agoYou can now run Inkling-Small, a new 276B model by Thinking Machines. Inkling-Small is the strongest open model for its size and runs local on 128GB RAM. Apache-2.0 Licensed, it has image, audio + 1M context support. Guide: unsloth.ai/docs/models/... GGUF: huggingface.co/unsloth/Inkl...
Interesting to see the slow meltdown in South Korea Granted it was a heavily leveraged market, but I'm curious to see if it will somehow trigger the pop of the bubble (Yes, I do think it's a bubble – I think AI is useful and will stay, y'all know me, but current financials make zero sense imo)
Surprised the talk about Jevons Paradox (which for some reason is not called Jevons' Paradox and it's also not a paradox but whatever) hasn't started ages ago www.youtube.com/watch?v=Vv3C...
Sam Altman - How to Start a Startup
YouTube video by Relentless
youtube.com
k, so Opus 5 is better than Fable 5? Why can't we semver properly?
What's a tech you are excited about that has not plateaued/turned to tiny incremental updates in the last decade? Let's ignore AI/LLMs for a sec.
Wanted to buy the Xteink X4 Pro But they went with pogo pins for charging instead of USB-C 🤦🏻♂️ That's an instant no from me, sorry...
I’m very very very impressed by the capital efficiency of the Chinese labs. Imagine what they could do with access to the amount of capital and GPUs that USA companies have.
Another one dropping soon The Chinese labs are on a roll
Asa@asa.engineer · 2w agoQwen3.8 (open weights) is coming soon
I’m selling 3 GPUs: - 1 x NVIDIA GTX 5060 Ti (16GB VRAM) - 2 x AMD Instinct MI100 (32GB VRAM each) I’m keeping the 5090. How long do y’all recon it will take?
I mean, usually it takes Open Weight models a few months to catch up... This is much faster than I expected
mr. TIM@timkellogg.me · 3w agothis wasn’t supposed to happen yet
Seeing rumors that Kimi 3 was released but can’t find it anywhere
@flashes.blue is there a way to specify which images I want to appear on my feed in your app? I do photography but also upload a bunch of random screenshots and things on here that I don't want to show up on my Flashes feed. This is pretty much why I've not been using the app so far.
New Open-Weights LLM dropped, openly designed to be fine-tuned. Love seeing this. 1T params and 1M context, so VRAM will be your main problem... huggingface.co/blog/thinkin...
Welcome Inkling by Thinking Machines
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
Codex encrypting subagent prompts I don’t like not seeing what my LLM is doing I own those subagent prompts, why are you hiding them from me? Switching to pi.dev or OpenCode from now on
github.com
After some time with both Sol and Fable: Sol is the better coder, it just keeps going and is less likely to miss stuff or introduce bugs Fable is the better researcher, it finds information usually more reliably than Sol Sol seems to be cheaper in terms of tokens used (haven't checked €/1M token)
Sure buddy. In the mean time y'all stay out of my 401(k) lol
Benn Jordan@bennjordan.bsky.social · 4w agoI found a hot stock tip for you.
