chad

@andthatto One thing I am noticing is that enabling `preserve_thinking` alone cuts down on thinking-death-spiral mode. I haven't seen it happen since enabling it. Watching the thought logs, thinking has become incremental, replies faster/more natural.

@chaddotphp nice!

Qwen 3.6 is frontier for local. It also thinks forever. I tried a dumb inference-time trick: make its block obey a tiny grammar. Result: - HumanEval+: 22x fewer think tokens, no accuracy loss - LiveCodeBench public slice: +14% pass@1, ~5x fewer total tokens

@andthatto Only thing is that doing this drops tool calling support on llama-server. The llama community is asking for a feature that would unlock using grammar with tools. I have it working locally so will open a PR after more testing if someone else doesn’t get to it first.

@bensig I'm sorry what

My friend Milla Jovovich and I spent months creating an AI memory system with Claude. It just posted a perfect score on the standard benchmark - beating every product in the space, free or paid. It's called MemPalace, and it works nothing like anything else out there. Instead of sending your data to a background agent in the cloud, it mines your conversations locally and organizes them into a palace - a structured architecture with wings, halls, and rooms that mirrors how human memory actually works. Here is what that gets you: → Your AI knows who you are before you type a single word - family, projects, preferences, loaded in ~120 tokens → Palace architecture organizes memories by domain and type - not a flat list of facts, a navigable structure → Semantic search across months of conversations finds the answer in position 1 or 2 → AAAK compression fits your entire life context into 120 tokens - 30x lossless compression any LLM reads natively → Contradiction detection catches wrong names, wrong pronouns, wrong ages before you ever see them The benchmarks: 100% recall on LongMemEval — first perfect score ever recorded. 500/500 questions. Every question type at 100%. 92.9% on ConvoMem — more than 2x Mem0's score. 100% on LoCoMo — every multi-hop reasoning category, including temporal inference which stumps most systems. No API key. No cloud. No subscription. One dependency. Runs on your machine. Your memories never leave. MIT License. 100% Open Source. github.com/milla-jovovich…

@ganyicz Lint for complexity and enforce a grade A rating

So, have we all decided to just accept this?

me too, little dude…

With every technological advancement we become a little more powerful and a little less essentially human

I wish I could talk to someone at Google and tell them how dumb Gemini has gotten in the past month. Getting cached responses to new questions a lot.

@shl doot

Reply for a Sora invite

@sentdefender and that’s probably why we want it back

Bagram Airbase was also originally built by the Soviet Union in the 1950s, not the United States.

@sentdefender Where’s all this bluster with Putin’s planes buzzing through NATO airspace?

U.S. President Donald J. Trump is continuing his spree of threatening posts today on Truth Social, stating now that if the Taliban does not give Bagram Airbase back to the United States, “BAD THINGS ARE GOING TO HAPPEN!!!”

@spaula6563 This website is unusable on mobile

what is bro doing in poland

@AutismCapital please no my house is full

We’re at the point of the administration where you just spin the Wheel of Deportation and hope for the best 😂

Billionaires Fight Club Vol.2 — Help Us Get Zuck & @sama to Repost 🚀 We made this fight because we love The Matrix. It’s probably the greatest film in history about AI. And it’s simply beautiful. We deeply respect both leaders and their impact on humanity. But someone had to play Agent Smith 😉 Sam, we don’t see you as a supervillain — in today’s world, you’ve simply become synonymous with “AI”, and someone had to step into that role for this piece to happen. Mark, we know how you love martial arts and once dressed as John Wick for Halloween… and honestly, who wouldn’t want to be Neo? 🥷 Community — we need your help. We love you, and we love the kind of movements where people come together and make the impossible possible. Let’s make this reality together — while we, like Neo, start to believe. ❤️

@shidoxo In an age where hiring is increasingly automated, the unfortunate optimization is what us old guys called keyword stuffing.

@Reagent_Systems @Stbatya @birdabo Is this unique to NASA? Is it because it’s hard to accurately represent real-world exit conditions?

@Stbatya @birdabo This is most likely the truest response. Just like someone else said, infinite loops could happen.

just learned NASA doesn’t allow recursion in their code.

@politicalmath Someday a company will make avatars of lost loved ones that subtly manipulate people towards decisions or beliefs that align with said company’s interests.

No, it's not. It's false. It's fake. The animated photos are untrue. It's just a small lie but it is still a lie. Your dead father never acted that way, never moved that way, never did that thing. We need to reject the small lies because they make way for larger lies

@forallcurious Bring snacks

if you could send a message to an alien civilization, what would you say?

@ailton I’m so glad we’re getting this instead of a functioning Siri

Liquid Glass merging is so beautiful #ios26 #liquidglass