Sources: several tech companies, including OpenAI, are encouraging the DOD behind the scenes to back away from designating Anthropic a “supply chain risk”
https://www.nytimes.com/2026/03/18/technology/silicon-valley-anthropic-pentagon.html?unlocked_article_code=1.UFA.qqG1.ysIXJtMWH4oP&smid=nytcore-ios-share
The seemingly friendly Silicon Valley lobbying to shield Anthropic from Pentagon scrutiny exposes a dangerous conflation of national security and corporate interests. OpenAI and others are effectively weaponizing their influence to neuter legitimate supply chain risk assessments, undermining transparency in an industry critical to geopolitical tech dominance. This behind-the-scenes push signals that the real threat may come not from the technology itself, but from the gatekeepers manipulating regulatory frameworks to protect market share at the expense of national security.
Show HN: Duplicate 3 layers in a 24B LLM, logical deduction .22→.76. No training
https://github.com/alainnothere/llm-circuit-finder
The demonstration of massive logical deduction improvement by duplicating layers without retraining suggests we are nowhere near optimizing large language models structurally. Yet, mainstream AI narratives hype incremental tweaks as breakthroughs, ignoring that the architecture’s brittleness leaves vast performance gains on the table. This experiment exposes how shallow current models remain and underscores the absurdity of overhyping “training data” volume without rethinking foundational design—an inconvenient truth for entrenched AI giants.
Google co-founder Sergey Brin increases his total contribution to $45M for the Building a Better California Super PAC that opposes a proposed 5% billionaire tax
https://www.theguardian.com/us-news/2026/mar/18/google-sergey-brin-california-billionaire-tax
Brin’s heavy funding of anti-tax lobbying is a stark reminder of how tech elites shape policy to entrench wealth, not innovation. Under the guise of “building a better California,” this PAC’s real agenda is protecting billionaire hoarding that exacerbates inequality and hampers public investment in tech infrastructure. The tech industry’s “self-regulation” narrative crumbles under such blatant influence-peddling, revealing that Silicon Valley’s political muscle is focused on rent-seeking, not societal progress.
Compute startup Andromeda raised funding from Paradigm at a $1.5B valuation, bringing Paradigm’s total to $60M, and passed a $100M revenue run rate in 2025
https://www.upstartsmedia.com/p/andromeda-ai-compute-startup-raises-60m
While Andromeda’s scaling suggests strong demand for AI compute, the valuation and revenue trajectory highlight the reckless capital exuberance fueling the compute arms race. The rush to dominate hardware layers risks creating monocultures vulnerable to supply chain shocks and geopolitical disruptions, especially given the global chip shortage and US-China tensions. Investors betting big on compute startups may be ignoring that AI’s future hinges as much on diversified strategy as raw processing power.
Ramp data: Anthropic is capturing ~73% of all spending among companies buying AI tools for the first time, up from a 50/50 split with OpenAI in January
https://www.axios.com/2026/03/18/ai-enterprise-revenue-anthropic-openai
Anthropic’s sudden dominance in first-time AI enterprise spending reveals a volatile market highly sensitive to brand positioning and perceived trustworthiness. Yet, this shift raises red flags about market concentration and overreliance on a single provider, increasing systemic risk if Anthropic faces disruption or policy backlash. The narrative of healthy AI competition masks the dangerous consolidation within an industry whose supply chains and data governance remain murky at best.
Nvidia CEO heralds ‘inference inflection’ as next phase of AI boom, backed by $1 trillion in orders
https://apnews.com/article/nvidia-ceo-jensen-huang-artificial-intelligence-conference-846f7d4aada068e92516665c6993ea29
Nvidia’s $1 trillion order pipeline signals a frenzy around AI inference but also exposes a looming bottleneck: the dependency on a single company’s hardware ecosystem. This “inference inflection” hype glosses over the geopolitical vulnerabilities inherent in supply chains dominated by concentrated chipmakers, especially amid rising US-China tech hostilities. The AI gold rush is less a sign of sustainable innovation than an arms race fueled by oligopolistic control, setting the stage for future market and geopolitical shocks.
Sources: Hacker News, Techmeme, AP News, Ars Technica | Compiled 2026-03-19