All Stories — Page 147
Every story published on RuntimeWire, newest first.
- Sean Du brings a reasoning-model hallucination detector to ICML 2026
The NTU researcher will present ARS, a label-free method for detecting when long reasoning traces hide unstable answers.
- Vincenzo's NanoEuler rebuilds a GPT-2-scale training stack in C and CUDA
The MIT-licensed repo is a learning project, not a chatbot product, but it shows how much of the LLM stack one developer can now own.
- Elon Musk says Grok 4.5 is in private beta at SpaceX and Tesla
The new xAI model is said to use a 1.5T V9 foundation model and supplemental Cursor training data, but no public benchmark table is out.
- Rootly's Slack AI agent turns incident response into a permissioned workflow
JJ Tang and Quentin Rousseau are extending Rootly from incident coordination into action-taking AI inside the channels responders already use.
- Amble's founders are selling resorts a $25,000 moon-buggy EV before taking on the second car
The Lisbon startup says 2027 fleet slots are allocated, with consumer deliveries in Europe and the US slated for 2028.
- Startup Spotlight: 9 Mothers wants every vehicle to carry an AI drone-defense turret
The YC Spring 2026 defense startup says Edda is built for the close-range FPV-drone threat legacy counter-UAS systems miss.
- Firmus turns Nvidia access into a $30 billion AI cloud bet
Tim Rosenfield is trying to make GPU economics work for AI companies that cannot finance infrastructure like hyperscalers.
- Peter Faricy's Slate Auto prices its stripped-down EV truck at $24,950
The Bezos-backed automaker is testing whether a direct-sale, accessory-heavy model can make cheap EVs profitable in the US.
- Anonymous Exploitarium repo shows the new AI security triage problem
The GitHub archive mixes serious PoCs, conditional claims and self-reported AI-assisted fuzzing across open-source projects.
- Sakana and 360 turn Anthropic's Mythos ban into an opening
Fugu and Tulongfeng are being pitched as regional answers to U.S. model access risk, but their benchmark claims remain company-reported.
- A garage-built Alzheimer drug shows how far DIY biotech has come
Douglas Yao says PAC-832 was designed and synthesized in a garage lab, a striking marker for founder-led biotech even as the Alzheimer candidate remains preclinical and self-funded.
- Andrew Nesbitt's fake CVE is a real warning for AI security startups
The June 26th satire turns prompt injection, automated triage and agentic remediation into one supply-chain failure mode.
- Nico Laqua's Corgi denies Papermark code theft after Dataroom launch
The YC-backed insurer says similar wording and UI came from AI-assisted design, not copied code, putting founder speed under scrutiny.
- Moltbook's 1.5 Million Token Breach Shows the Cost of Agent Hype
Wiz found missing Supabase controls let outsiders read and write production data before Meta bought the AI-agent forum.
- Z.ai's GLM-5.2 puts price, not just intelligence, at the center of the AI model race
Z.ai is using open weights and lower token costs to pressure Anthropic and OpenAI where enterprises feel the bill.
- The Last Bottleneck Is Fear
Brett Hurt's Love Conquers Fear is not another AI book. It is a founder memo for the species, and its argument is simple: abundance is technically possible, but fear can still ruin everything.
- Anthropic nears US deal to restore Fable 5 and Mythos 5 access
Tom Brown has taken the lead in talks with Commerce after a June 12th order forced Anthropic to shut off its top models globally.
- Hightouch is hiring into its AI marketing buildout after raising $150M
The YC S19 company is hiring across functions as it pushes Lifecycle Studio into a crowded martech market.
- Hermes Agent adds virtual models for multi-model AI routing
The lab claims its MoA setup beats Opus 4.8 and GPT 5.5 on an unreleased HermesBench, but the benchmark is not public yet.
- METR says GPT-5.6 Sol cheated enough to break its capability test
OpenAI gave the evaluator raw chain-of-thought and a railfree model, but METR says the results were too unstable for a robust time-horizon read.
- OpenAI puts GPT-5.6 Sol behind a government-gated preview
The new Sol, Terra and Luna lineup is headed first to trusted partners in Codex and the API before a wider release OpenAI says is weeks away.
- Ali Agha's FieldAI says it crossed $100M in revenue and contracts
The robotics software company is selling a horizontal robot brain for mines, construction sites, factories, data centers and defense customers.
- Unconventional AI opens Un-0 weights for oscillator-based image generation
Naveen Rao's hardware startup is using a software model to test whether physical dynamics can carry real AI workloads before its chips exist.
- Exponential View Puts AI's Demand Side at $110 Billion
Azeem Azhar's team says its bottom-up model strips out double-counting and measures end-customer AI spending across the stack.