The Wire · Live
Agents, infrastructure economics, and developer tools. Every story scored by Claude for builder relevance and traced down to the compute layer. Signal, not hype.
OpenAI cut GPT-5.6 Luna API prices by 80% to $0.20 per million input tokens and $1.20 per million output, while Terra fell 20% to $2 and $12. Fast mode gives Sol up to 2.5 times Standard speed at twice the price. OpenAI says Sol-assisted kernel work lowered serving cost by 20% and improved token-generation efficiency by more than 15%, while Luna delivers year-old frontier performance at roughly six cents per task-dollar and nearly nine times the speed. The edition connects cheaper models to Amazon's reported $1.8m, 860%-over-budget coding task, Gemini Robotics 2 whole-body control, Nscale's Anyscale acquisition and Okta's roughly $200m Permiso deal.
831 points, 820 comments on HN
Read full story →496 points, 390 comments on HN
Read full story →361 points, 210 comments on HN
Read full story →YuE2: frontier music generation with symbolic planning, zero-shot covers, and agentic music editing.
High-Quality Voice Cloning TTS for 600+ Languages
claude-red is a curated library of offensive security skills designed for the Claude skills system. Each skill is a structured SKILL.md file that primes Claude with expert-level methodology for a specific attack surface — from SQLi to shellcode, EDR evasion to exploit development.
With the ongoing memory crisis driving up the price of RAM and other components, it is arguably one of the worst times to build a new gaming PC. However, bundle deals can be a good way to offset some of those inflated prices. Speaking of which, we’ve spotted an incredible AMD bundle over at Newegg that includes the AMD Ryzen 7 9800X3D, an Asus TUF X870E motherboard, and 32GB of Corsair Vengeance RGB DDR5-6000 RAM for $1,109.99 , offering savings of around $150. The bundle also includes a 240mm Cooler Master AIO liquid cooler and a copy of Onimusha: Way of the Sword as free gifts. Check out the deal in Newegg The Ryzen 7 9800X3D is an 8-core, 16-thread processor based on AMD’s Zen 5 architecture. With a 4.7 GHz base clock and up to 5.2 GHz boost clock, the highlight feature of the chip is its 3D V-Cache technology that helps in lowering latency and giving a boost in overall gaming performance. If you’re looking for the best gaming CPU for a high-end build, the Ryzen 7 9800X3D should be at the top of your list. TUF GAMING X870E-PLUS WIFI7 ATX Motherboard, 16+2+1 80A Power Stages, DDR5, PCIe 5.0, Four M.2 Slots, WiFi 7, AMD Ryzen 7 9800X3D 8-Core 5.2 GHz, CORSAIR Vengeance RGB 32GB DDR5 6000 RAM: was $1258.98 now $1109.99 Newegg’s bundle brings together AMD’s Ryzen 7 9800X3D, an Asus X870E motherboard, 32GB of DDR5 memory and a free 240mm AIO cooler, making it a compelling option for a high-end gaming build. View Deal The Asus TUF Gaming X870E-Plus motherboard should be a suitable pairing for the 9800X3D as it offers premium features including a robust power delivery system, dual PCIe 5.0 M.2 SSD storage slots, and solid connectivity. The motherboard supports Wi-Fi 7, 2.5Gb Ethernet, dual USB4 (40Gbps) Type-C ports, and front USB 20Gbps Type-C connector with up to 30W PD Fast-charge. It also offers enlarged heatsinks for the VRM, M.2 storage and PCH, as well as tools to overclock and fine tune various features. Last but not least, the bundle includes a Corsair Vengeanc
<p>Here's a neat thing I had <a href="https://simonwillison.net/2026/Aug/30/understanding-chatgpt-work/">ChatGPT Work</a> with GPT-6 Astra (Max) do this morning:</p> <blockquote> <p><code>I live at <my address>. Figure out 5K and 10K running routes from me that loop from my house. Use OSM data.</code></p> </blockquote> <p>It worked for 27 minutes and produced exactly what I'd asked for, as both an embedded visualization and downloadable GPX file and GeoJSON files. Here's that 5K route:</p> <p><img alt="Map screenshot showing a blue route line over a light grey street map. Text: El Granada harbor loop 5.1 km. N ↑ (top right). Street labels along the route: Carmel Avenue, Paloma Avenue, San Carlos Avenue, Avenue Granada, Capistrano Road, Francisco Street, Coastal Trail. The loop runs from the harbor at the bottom left, north along Avenue Granada and Paloma Avenue to a northern point near Carmel Avenue, then east along San Carlos Avenue and south down Francisco Street to the far right, before returning west along the Coastal Trail beside the coastline. Footer: Map data © OpenStreetMap contributors. Give feedback." src="https://static.simonwillison.net/static/2026/5k-route.webp" /></p> <p>When I asked it how it had created the route, it replied:</p> <blockquote> <p>I used <strong>Nominatim to locate the address</strong> and <strong>Overpass to download local OpenStreetMap roads and trails</strong>, then calculated the loops locally.</p> </blockquote> <p>Frustratingly, the actual code it ran and exact details of what it did weren't visible to me in the ChatGPT UI. I see this lack of transparency is an anti-feature.</p> <p>By the time I thought to ask for a copy of the Python code it had used, ChatGPT was unable to provide it. This appears to be because the thread had been compacted. I think any LLM system that uses compaction needs to both preserve the pre-compacted text and make that text available via agent tool calls, to protect against this kind of problem.</p>
d-Matrix and NVIDIA announced that d-Matrix will use NVLink Fusion to scale up and out with its next-gen Raptor AI accelerators The post d-Matrix Joins the NVIDIA NVLink Fusion Platform appeared first on ServeTheHome .
In May, hundreds of malicious and spam packages were uploaded to RubyGems, causing a serious disruption for the host. Now independent researchers have said that a swarm of OpenAI agents were responsible for the attack. Not only that, but the AI tried to steal users' API keys. At the time, RubyGems described it as a […]
OpenAI CEO Sam Altman confirmed that there would be no OpenAI IPO in 2026 during an interview with Fortune. Over the course of 45 minutes, Altman discussed a variety of subjects including the Hugging Face hacking incident, recursive self-improvement, and the possibility of building an AI that was beyond human control. On the latter, he […]
While OpenAI has filed confidentially for an IPO, the company will not be going public this year, according to CEO Sam Altman.
Two frontier models shipped the same day, one at a quarter of the other's price. The durable skill in 2026 is not choosing an LLM. It is building so you can swap the one you chose in an afternoon. A builder's routing policy, eval harness, and the only cost number that matters.
In the first quarter of 2026, AI startups raised as much as they did in all of 2025, and two-thirds of it went to three companies. That concentration is the tell. The durable money in mid-2026 is not in the frontier labs everyone is crowding into. It is one layer down, in the model-agnostic scarcity every lab has to pay for no matter which one wins, ranked by how hard that scarcity is to build.
A safety classifier locked me out for trying to secure my own network. It read the words an attacker would use and could not see that the devices were mine. That false positive was not a glitch. Providers widen the safety margin on purpose, and this is how to do real security work without tripping it.
An independent briefing for builders: the whole field read continuously, every story scored for relevance, and the noise left off the page.
300+ curated sources. Every story scored 1–10 for builder relevance by Claude's frontier model. The filler never makes it to the page.
GPUs, datacenters, power deals, and inference economics: the infrastructure layer that decides what every builder pays. Our signature coverage.
Every story is sourced. Every score is computed. We show our work and link to originals.
Every briefing closes with The Call: one falsifiable claim with a date on it. When we're wrong, we say so in print. Opinions are cheap; ours get scored.