GPT-2 in Pure CMake: Because Why Not
AlpinDale's gpt2.cmake implements a GPT-2 model entirely in CMake, using Q16.16 fixed-point arithmetic. It's a wild proof-of-concept that pushes the build tool far beyond its intended use.
Aug 24, 2026· GitHub
News
News
Tech news, summarized and translated.
Google Workspace Blocks Legit Domains with a Broken Regex Blocklist
A developer discovers Google Workspace's sign-up page rejects valid domains like web.example.com because of a hardcoded regex list of 'email providers' — and the fix is a frontend-only bypass.
Aug 23, 2026
Agent Harnesses: The Software Layer That Lets You Own Your AI
An agent harness is the software shell that turns a raw AI model into a tool you can actually control — and it's the key to breaking free from AI lab lock-in.
Aug 23, 2026· Earendil
Why Your Local LLM Feels Dumber Than It Is: Attention Backends Quietly Change Outputs
A deep-dive experiment shows that swapping attention backends in vLLM can flip the model's top-1 token choices on real agentic workloads, even with identical weights and hardware. The culprit: implementation-specific math, not the model.
Aug 23, 2026· Level1Techs Forums
Atproto Spaces Alpha: Private Data on a Public Network
The AT Protocol's biggest update yet introduces Spaces, a new primitive for gated, non-public data with access control (not encryption). Alpha code, SDKs, and a hosted PDS are live for experimentation.
Aug 22, 2026· AT Protocol
Qwen3-TTS at 10 RPS with sub-50 ms TTFA: Nari Labs pushes the speed-cost frontier
Nari Labs claims a Qwen3-TTS 1.7B CustomVoice serving stack hits 10 RPS with sub-50 ms p95 time-to-first-audio on a single H100, beating vLLM-Omni, SGLang-Omni, VoxServe, and M* after tuning. The key: a unified scheduler for Talker, Code Predictor, and Codec.
Aug 21, 2026· Nari Labs
Photoshop on a 60p Chip: The Case for Simpler Computers
A developer runs Photoshop on a Raspberry Pi RP2350 chip that costs 60 pence, emulating a 1990s Mac. It's a reminder that modern computing power is absurdly cheap, and a nudge to question the bloat we accept.
Aug 21, 2026
Spring AI 2.0.1: Seven CVEs Fixed, Tool Call Limits, and Streaming Reliability
Spring AI 2.0.1 is a hefty first maintenance release for the 2.0 line: seven security fixes, a new tool-call limit to stop runaway agent loops, OpenAI audio streaming, and a pile of streaming/tool-calling reliability patches.
Aug 21, 2026· Spring AI 2.0.1 Available Now
Spring Boot 4.1.1 ships with 98 fixes and dependency upgrades
Spring Boot 4.1.1 is out on Maven Central, bringing 98 bug fixes, documentation tweaks, and dependency upgrades. A routine patch release for teams already on 4.1.x.
Aug 21, 2026· Spring Boot 4.1.1 available now
Cloudflare OAuth Gets Task-Based Consent: Users Can Now Opt Out of Optional Scopes
Cloudflare introduces OAuth scope customization, letting users grant narrower access at authorization time. Developers mark scopes as optional, and the consent screen adapts — a practical step toward least-privilege for agents and MCP servers.
Aug 21, 2026· Cloudflare Blog
Spring Boot 4.0.8 ships with 77 fixes and dependency upgrades
Spring Boot 4.0.8 is out on Maven Central, packing 77 bug fixes, documentation improvements, and dependency upgrades. A routine patch release for teams on the 4.0 line.
Aug 21, 2026· Spring Boot 4.0.8 available now
Spring Cloud 2025.1.3 (Oakwood) Ships with a Stack of Security Fixes
The latest Spring Cloud release train, 2025.1.3 (Oakwood), is out with a heavy focus on security—patching multiple CVEs across Config, Function, Gateway, Stream, and Commons—plus a few functional improvements.
Aug 21, 2026· Spring Cloud 2025.1.3 (aka Oakwood) Has Been Released