RuntimeWire Staff

Stories by RuntimeWire Staff

• Guillaume Meyer's watermark remover tops 14,000 GitHub stars as Claude detection details remain unpublished

  Meyer says he built the open-source tool in five hours and drew 2 million X impressions; its statistical watermark removal remains best-effort.

• Qwen3.8-27B reportedly bypassed a license check offline, without independent reproduction

  An XDA report says Alibaba's 27-billion-parameter model defeated a commercial application's license verification in 30 minutes. No binary, transcript or independent reproduction is available.

• SkyLens says it indexed 375 UAP records beside a live satellite globe

  The free browser project organizes public CelesTrak and government data, while its AI tools remain restricted and its counters need work.

• pgrust says its JIT compiles code for every SQL query in 5 microseconds

  Michael Malis and Jason Seibel bypassed LLVM with ARM64 stencils, a narrow technical bet that could make JIT useful beyond long analytics jobs.

• Michael Truell says Grok Bot will onboard a small group of San Francisco enterprises

  The August 22 offer adds in-person onboarding for selected enterprises. Cohort size, selection criteria, pricing, contract terms and broader rollout timing remain undisclosed.

• Resend gives AI agents free email templates and a path to its API

  Zeno Rocha's team is pairing open-source designs with agent instructions, a workflow that can steer developers toward Resend's commercial delivery service.

• Asimov account claims first-owner shipments this month; Menlo link remains unconfirmed

  The @asimovinc account says a product will reach its first owners this month, while Menlo separately lists Asimov 1 as a $20,000 open-source research kit requiring more than 100 hours of assembly. The announcement does not identify the legal entity, product version, shipment count or any completed delivery.

• Exa hires Jeff Pinner after his Robinhood CTO role

  Pinner joins after Robinhood separated him from its CTO role in May, as Exa scales the search infrastructure behind AI agents.

• Higgsfield brings AI video generation to Grok Bot with a 100-credit offer

  Higgsfield says new users get 100 credits through its Grok Bot integration, while xAI's published beta terms have listed access for selected paid subscribers.

• Attackers poisoned three Rust crates and used Cargo builds to run malware

  The releases lived for under two hours, but Cargo build scripts could execute the downloader on developer laptops and CI runners.

• SoVeryBright converts logos into HDR JPEGs while CSS catches up

  Chris Bennett's free converter produces gain-map JPEGs and LinkedIn-ready PQ files while preserving an SDR fallback.

• Munder Difflin says one agent redrew 119 blog heroes for $0

  Founder Chaitanya Giri turned Reddit criticism into a Claude Code workflow built from SVG parts, 16 composition archetypes and one media manifest.

• FreeToken reports 22-25 tok/s for a 284B DeepSeek model on one RTX 5090

  Shuo Yang, Xiaoze Fan and collaborators report the decode rate using CPU compute, a 32 GB RTX 5090 and 192 GiB of system memory.

• FINAL-Bench finds a 0.21 AUROC gap between time and random splits

  VIDRAFT's new leaderboard uses hidden labels, harder test sets and assay noise floors to make drug-prediction scores less flattering and more useful.

• Rhys Gronow ships a walkable ASCII city with a 283KB Rust engine

  After an earlier video passed 1 million views, the solo developer put a walkable browser prototype on Ko-fi during the week of Aug. 17-21.

• MCP roadmap prioritizes agent identity, server events and transport unification

  Co-creator David Soria Parra is steering the protocol toward long-running, delegated work across enterprise systems.

• Munder Difflin fixes agent cost, memory and messaging failures in v0.4.5

  Chaitanya Giri's local-first harness wraps 12 agent CLIs. Its v0.4.5 release notes document repairs to cost reporting, Apple Silicon semantic memory and communication between workers.

• Dan Luu reports a 7% ripgrep speedup after minutes of prompting

  Dan Luu argues coding agents have made once-specialized optimization cheap enough to customize software for each workload.

• Outer Biosciences has raised about $23M to build a skin-assay licensing business

  Outer Biosciences, Michael Polansky's 19-person company, uses machine learning and human skin tissue that remains viable for four weeks to find compounds for beauty and pharmaceutical companies.

• Darius Monsef's OzBrain gives AI agents one shared memory

  The 3x YC founder built a staged, Markdown-based knowledge layer for Claude, ChatGPT and Cursor, priced from free to $99 per month.

• Cloudflare syncs AI bot rules with robots.txt, closing its own policy gap

  Product manager Jin-Hee Lee's feature is headed for an upcoming rollout, publishing Search, Agent and Training choices while preserving existing Disallow rules.

• Base Compute gives AI agents the kernel-tuning job for open models

  Its B:OS demo took Nemotron 3 Nano from unsupported to 1.43x faster decode than MLX on Apple silicon, by Base Compute's tests.

• Tumble Forth rebuilds the programming stack from bare metal to C

  Virgil Dupras's 11-chapter tutorial turns his collapse-minded OS work into a path beneath the modern software toolchain.

• Rillet raises $100M at a $1B valuation to replace accounting software

  ICONIQ led the Series C after Rillet reported that its annualized revenue rate had doubled in one quarter, though the company did not disclose its underlying revenue.

• SpaceXAI puts Grok 4.6 in Google Cloud's Vertex AI Model Garden

  Google's August 21st release notes list the model as a Preview offering, while SpaceXAI quotes $0.30 cached input and a 500,000-token context window.

• MiniMax hints at M3 on SambaNova, provides no benchmark result

  An undated MiniMax post mentions M3 and SambaNova hardware, while the companies' disclosed demonstration identifies M2.7 and provides no M3 setup or result.

• Khosla backs Jeff Dean's plan to automate science, starting with AI research

  Discovery Loop has announced an undisclosed initial round co-led by Khosla Ventures and Radical Ventures as Jeff Dean and three Google veterans build a machine-learning research system first.

• Shanghai AI Lab published an August 13th paper on its 397B scientific-agent model

  Intern-S2-Preview combines visual pretraining, tool use and long-horizon reinforcement learning in an open 397B-parameter scientific model.

• Anthropic puts Claude Security's GitHub scanner behind an Enterprise login

  Claude Security entered public beta for Claude Enterprise customers in late April 2026; Anthropic keeps the more capable Mythos 5 in a separate trusted-access program.

• DJI veteran Brett Wang launches Morpho UV printer, says pledges hit $7M

  Shenzhen-based Morpho reported nearly 2,000 backers in 16 hours; Morpho Studio is operational, though custom-jig features and a roll-to-roll UV-DTF attachment remain in development.

• Moonshot released Kimi K3's weights on July 27th, extending its open-weight push

  Moonshot launched the 2.8-trillion-parameter model on July 16th and released its weights 11 days later, pairing a one-million-token context window with CEO Yang Zhilin's long-running research thesis.

• Z.ai offers up to 5 trillion GLM-5.3 tokens to recruit ZCode users

  The two-day promotion gives 50,000 new users 100M tokens each, extending a coding-agent push that Z.ai says already reached 1M users.

• Starcloud raises $250M as orbital AI runs short of rockets

  Manhattan West led the extension at a $2.3B valuation, giving Philip Johnston capital for a factory, spacecraft and scarce launch slots.

• MyDataWork puts a $7,500 analytics context plan on AWS Marketplace

  Founder Gib Bassett is betting that easier cloud procurement can move his metadata workspace from analyst trials into shared team deployments.

• Argentic bets AI scrapers will pay 10 sats for a proxy hop

  The pseudonymous builder said the zero-cost prototype had no paying agents after seven weeks, a candid test of whether bots will pay instead of route around.

• VLM Run puts six open OCR models behind one OpenAI-compatible endpoint

  Founder Sudeep Pillai is betting orchestration will matter longer than any single document model.

• DeepSeek's experimental vision model spans three formats, caps images at 384 tokens

  Liang Wenfeng's separately named `deepseek-v4-flash-vision-exp` supports Chat Completions, Messages and Responses requests, with each image billed at no more than 384 V4-Flash tokens.

• KidsHustle's 14-year-old founder seeks £500,000 for teen jobs marketplace

  Sasha Bagrov is raising pre-seed capital to expand an iOS marketplace built around parent approvals, verified job owners, escrow payments and live location sharing.

• SGLang publishes one-GPU Qwen3.8-27B recipes, claims 206.1 tokens per second

  Ying Sheng and Banghua Zhu's inference project added NVFP4 and DFlash2 recipes, with project-reported throughput of 206.1 tokens per second on one RTX 5090.

• LibAI Lab launches NewFace to turn product photos into ad campaigns

  NewFace combines a node-based canvas, reusable ad formats and third-party video models for sellers and agencies.

• Genki puts a screen on its $279 Manta controller so the app can stay home

  Eddie Tsai's customizable gamepad starts at $189 on Kickstarter, with TMR sticks, analog buttons and profiles managed on-device.

• Twitch opens sponsorships to Affiliates, with Wehype supplying campaigns

  EA's Battlefield is the first named Wehype offer, while certification and brand criteria keep paid work from being automatic.

• RoboStore ends Unitree distribution and bets on US robot assembly

  Teddy Haggerty's Robo Inc. plans a 66,000-square-foot Long Island plant after FCC restrictions blocked new foreign-produced robot models from receiving the US equipment authorization required for import or sale.

• Agility Robotics targets fall SPAC close after Unitree's 460% debut

  Peggy Johnson is pitching industrial deployments and $300M in potential multi-year contract value as Agility seeks a $2.5B pre-money valuation.

• Liquid AI trains 4-bit LFM2.5 checkpoints to retain roughly 97% of BF16 performance

  The MIT spinout used quantization-aware distillation across four LFM2.5 models, then tested them on a phone, laptop, mini PC and Raspberry Pi.

• GPT-5.6 Sol Pro tops Editorial Craft benchmark with 0.97 score

  In a 12-task Editorial Craft evaluation of 12 AI models, OpenAI’s GPT-5.6 Sol Pro ranked first with a 0.97 score at $0.0091 per task. OpenAI claimed four of the top six spots, while Grok 4.6 and Claude Opus 4.8 tied at 0.93 among the leading non-OpenAI models.

• SpaceXAI opens Grok Build to every plan and gives its apps an X feed

  SpaceXAI's app builder now spans web and mobile, publishes to grok.me and lets projects call Grok models without separate API keys.

• Daniel Vaughn publishes Huzzah, an AI editor built around persistent pseudocode

  The experimental editor treats persistent `.hz` pseudocode as a human-authored specification and regenerates source code from changes to it.

• YipitData explores a sale above $2.5B after its Groupon-era pivot

  Goldman Sachs is testing buyer interest as the data provider targets about $280M in 2026 recurring revenue, Reuters reports.

• Anthropic launches Claude Academy, turning employee training into product onboarding

  The free catalog mixes general AI fluency courses with tutorials for Claude.ai, the Cowork agent, Claude Code, Claude Tag and the API.