Nostr pub key: npub197swtr3ymz7aw72tu7eqspxjta7zwcwf23uw4fxyxe2r2eyrfa2sdpner3

🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced! 🔷 Improved coding & agent performance over K2.6: 21.8% on Kimi Code Bench v2, 11.0% on Program Bench, and 31.5% on MLS Bench Lite. 🔷 Reasoning efficiency: Less overthinking, with 30% lower reasoning-token usage compared to K2.6. 🔷 Long-horizon coding: Improved instruction following, higher end-to-end coding task success rates. ⚡️ 6x High-Speed Mode coming soon! 🔌 Available today via Kimi API and Kimi Code. 🔗 Kimi Code: kimi.com/code 🔗 API: platform.moonshot.ai

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time. I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

Huawei released KVarN, a KV-cache compression method for LLMs It delivers 3-5x more context length, beats FP16 throughput, and keeps FP16 accuracy. One flag in vLLM. Zero calibration.

Sparrow v2.5.2 has been released to strengthen PSBT verification on load and add dust detection for SP wallets, along with several smaller fixes. Changelog: github.com/sparrowwallet/spa…

Bisq v2.1.11 released. Main focus of this release: • Major security and hardening improvements • Stronger release verification and update protection • Better Tor reliability and macOS Apple Silicon support Please update as soon as possible!

We are making our discount permanent! 🎉 Enjoy building with DeepSeek-V4-Pro and bring your innovative ideas to life! 🚀

Sparrow v2.5.1 has been released to fix a UI issue where the wrong script type was selected on loading a wallet with a non-default type. In addition, BIP322 message signing has been adjusted to match the now completed BIP. Changelog: github.com/sparrowwallet/spa…

Sparrow v2.5.0 released with: Silent Payments (SP) wallets SP-capable Frigate public server @2140_dev bitview.space fee rates source sparrowwallet.com/download

Today, we’re launching the @link wallet for agents. It lets you securely empower agents to spend on your behalf. Your payment credentials are never exposed and you approve every purchase. link.com/agents

OpenClaw 2026.4.20 🦞 🧠 Kimi K2.6 support provider-aware /think 💬 BlueBubbles iMessage sends tapbacks fixed ⏰ Cron state/delivery cleanup 🔐 Gateway pairing plugin startup hardening Less haunted. More useful. github.com/openclaw/openclaw…

🚨 [ICYMI] extracting data from PDFs just got solved. Someone open-sourced a tool that turns PDFs into Markdown at 100 pages a second 🤯 It’s called OpenDataLoader. It runs flawlessly on CPU and decodes tables, complex layouts, and nested structures like an absolute pro. Best part? 100% free and open-source. Grab the repo link in the 🧵↓

We're delighted to announce that MiniMax M2.7 is now officially open source. With SOTA performance in SWE-Pro (56.22%) and Terminal Bench 2 (57.0%). You can find it on Hugging Face now. Enjoy!🤗 huggingface：huggingface.co/MiniMaxAI/Min… Blog: minimax.io/news/minimax-m27-… MiniMax API: platform.minimax.io/

🇮🇷 Iran is charging $2M per ship to cross the Strait of Hormuz and they want it in Bitcoin. 😳 At $72,000 per $BTC, each ship = 27.7 BTC. Pre-crisis, 130 ships crossed daily. • Daily: 3,611 BTC • Monthly: 108,333 BTC • Yearly: 1.3 million BTC The entire Bitcoin network only mines 450 BTC per day. Iran would accumulate 8x the monthly mining supply. Every month. A sanctioned nation building a Bitcoin treasury through a toll booth. This is the most important geopolitical Bitcoin story nobody is talking about. 🔥

BREAKING: 🇺🇸 US Treasury just did the largest Treasury buyback in HISTORY. Treasury bought back $15,000,000,000 of its own debt to improve liquidity.

OpenClaw 2026.3.31 🦞 🇨🇳 Bundled QQ Bot — private, group, and guild chat media 📹 LINE now sends images, video, and audio 🧵 Real background task flows: list, show, cancel 🇯🇵 Better CJK: context, memory, and TTS OpenClaw's next release has been leaked🦞github.com/openclaw/openclaw…

In case you don't realize what just happened: MARA sold 15,000 BTC, Bitdeer sold all of its BTC, Riot is selling BTC from treasury to fund data center build-outs, Auradine just rebranded to Velaura AI. The biggest miners are leaving the game. Not necessarily because it's broken, but because AI pays more per megawatt. Think about what this means: These aren't small players. They find thousands of blocks and move global hashrate. When they redirect capital and infrastructure toward AI, that hashrate comes offline. Unless equivalent hashrate fills the gap, difficulty drops. And it doesn't seem the the gap will be filled in the near term because the whole reason they're leaving is better margins elsewhere. Lower difficulty = higher margins for every miner who stays. And there's a second layer here. If the Strait of Hormuz stays closed into April, energy prices climb. Oil-dependent miners get hit hardest. This might be the best setup small/medium miners with stable PPA's have seen since the 2021 China mining ban.

Google dropped the TurboQuant paper yesterday morning. 36 hours later it's running in llama.cpp on Apple Silicon, faster than the baseline it replaces. the numbers: - 4.6x KV cache compression - 102% of q8_0 speed (yes, faster, smaller cache = less memory bandwidth) - PPL within 1.3% of baseline (verified, not vibes) the optimization journey: 739 > starting point (fp32 rotation) 1074 > fp16 WHT 1411 > half4 vectorized butterfly 2095 > graph-side rotation (the big one) 2747 > block-32 graph WHT. faster than q8_0. 3.72x speedup in one day. from a paper I read at dinner last night. what I learned along the way: - the paper's QJL residual stage is unnecessary. multiple implementations confirmed this independently - Metal silently falls back to CPU if you mess up shader includes. cost me hours - "coherent text" output means nothing. I shipped PPL 165 thinking it worked. always run perplexity - ggml stores column-major. C arrays are row-major. this will ruin your afternoon everything is open source. the code, the benchmarks, the speed investigation logs, the debugging pain, all of it. github.com/TheTom/turboquant… paper to parity in 36 hours. what a time to be alive.

Every week a new "build agents with zero code in 3 steps" tool launches. The building part was never hard. The hard part is the agent doing something useful on day 2. I've watched hundreds of skills get listed on Claw Mart — the ones that stick aren't the clever demos, they're the boring ones that solve one specific workflow someone actually has.

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI