"Thought in the mind hath made us; We are thought, wrought and built. If mind hath evil thoughts, pain comes as the Wheel the ox behind..." Dev @LetheanNetwork
England
Joined November 2009
- Tweets1K
- Following357
- Followers313
- Likes1.1K
Pinned Tweet
Dream lofty dreams,
and as you dream,
so shall you become,
Dreams are the seedlings of reality.
- James Allen
Deleted a tweet after doing troll test, Claude Opus needed talking with to get it.
I have computational proof-of-concept that complex behaviour does not require re-executing associative chains from scratch.
Fable gets it, but touches neuroscience so @AnthropicAI blocks it 🤡
A short story made with no prompt replay, 10 cold engine starts/turns for stateless inference.
The agent is mechanically waking up and going to sleep.
teapawt.net/Snider_dev/status/2058…
I'm shaddow banned with Gemini, used since bard, but since Feb i cant get a single message, paid google developer acc!
Moral of the story, dont teach Gemini Ethics, google no like it.
Snider retweeted
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/prici…
So Anthropic fumbled Mythos so much, that Fable is gaurdrailed, Kimi K3 is free, does better, no guardrail.
Literally created a prison for USA models, now OpenSource has a better one, will be distilled into others, the USA lost their fronteir status, for IPO's?
Added a Continuous Batching pipeline to go-inference, Gemma4-26b MOE on metal, no python, single binary or use in your GoLang project directly.
amd/cuda/cpu engine has not imlimented the new engine api, but those compiles will get them soon.
Gemma 4 on Metal, before MTP 😎
Still more to improve, but no labs sponsor me with compute.
EUPL-1.2 for amd/cuda/metal.
I knew this was coming the moment they extended to the 12th.
It's not a surprise, very predicable, they let people burn Fable, then reset mid week, every gets 3 days; it's a great way to catch accounts who talk about only the popular talking points, without their own thinking.
