Quick Hits - SiliconNoon (Page 4)

Quick Hits

Quick Hits — August 11, 2026

A midday batch with a capital-markets flavor: the analyst case for treating the AI buildout like the 1870s railroad boom, Shanghai's plan to weld domestic chips to mainstream models, and Kai-Fu Lee's Nobel prediction. Ben Thompson argues the AI buildout is starting to look like the 1870s railroad boom — with the risk quietly moving onto institutional investors. His new Stratechery essay, "Nvidia's Risky Business," opens with Jay Cooke's 1870 railroad-financing contract and draws the line to Nvi

Quick Hits

Quick Hits: August 11, 2026

A quiet overnight batch: Anthropic holds the line on API pricing, OpenAI loses its ethics chief, a model-customization startup's valuation keeps compounding, and the small-model crowd gets fresh data points on just how cheap the edge is getting. Anthropic made Claude Sonnet 5's introductory pricing permanent, canceling a planned September 1 increase. The model stays at $2 per million input tokens and $10 per million output tokens, scrapping the 50 percent hike the company had telegraphed for ne

Quick Hits

Quick Hits: August 10, 2026

Three quick ones from the past day: a startup betting AI agents can find cooler chips, a finance-focused agent claiming top marks on tax, and Beijing quietly standardizing AI in regulated industries. YC-backed Discovered Materials raised a $9 million seed round to hunt for new semiconductor materials with swarms of AI agents. The startup, led by Stanford materials-science PhD Akash Ramdas and agent veteran Advaith Sridhar, runs Anthropic models in a custom harness to generate material leads, th

Quick Hits

Quick Hits: August 9, 2026

Four quick ones from the past day: a GPU-free DeepSeek milestone on China's flagship supercomputer, the spread of local "AI bureaus," the state bank cashing in on China's AI listing boom, and DeepSeek buying a body for its AGI ambitions. A Chinese team says it has run the full DeepSeek-V3/R1 671B model on the Lingsheng supercomputer using only CPUs — not a single GPU. Qingcheng Jizhi, an inference-optimization startup, says it deployed the MoE model across 16 nodes of Lingsheng's LX2 CPU platfo

Quick Hits

Quick Hits: August 9, 2026

Three stories worth a minute of your day: the distrust economy around AI writing detectors, the countermeasures market growing up around AI wearables, and a new free tool for turning text into voice. AI detectors are creating a new era of distrust, The Verge's Emma Roth argues in a column published today. The tools that promise to flag AI-generated writing remain unreliable, yet educators and publishers are adopting them anyway — and Roth's case is that this is quietly poisoning classrooms and

Quick Hits

Quick Hits: August 8, 2026

Four things worth your time today: a Mars driving record built on autonomy, two llama.cpp upgrades for the big-model era, and the local-benchmark case for MoE. NASA's Perseverance is about to set the all-time off-world driving record — and roughly 90% of the mileage was self-driven. Sometime next week the rover will pass 45.16 km on Mars, breaking the record Opportunity held since 2018, and it did it in about a third of the time. Its onboard Vision Compute Element lets it image the terrain and

See all