Background & Credentials
About
Marcus Thompson has spent over a decade at the forefront of developer tooling, building the frameworks and infrastructure that power modern web applications. As a former Staff Engineer at Vercel, he was a core contributor to Next.js, helping shape the framework that now powers millions of websites worldwide.
His journey through the developer tools ecosystem includes a pivotal role at Shopify, where he led the architecture of the Hydrogen framework, enabling developers to build custom storefronts with unprecedented performance. Marcus brings this deep, hands-on experience to his coverage of the rapidly evolving developer tools landscape.
With an MS in Computer Science from MIT, Marcus combines academic rigor with practical engineering wisdom. He's passionate about developer experience and believes that great tools should be invisible, getting out of the way so developers can focus on building products that matter.
Articles by Marcus (20)
Nist AI Agent Standards Push, Identity and Rfi
Open protocols, identity rules, and hard dates. Standards could cut integration costs - or calcify the stack if they land too early.
Uspto Resets AI Inventorship, Humans Must Own Claims
AI still is not an inventor. USPTO scrapped its 2024 test, clarified Pannu, and drew hard lines on U.S. priority and Europe/PCT mechanics.
Fei-Fei Li's World Labs Raises $1B for 3d World Models
Autodesk drops $200M into a $1B round as World Labs scales Marble and a World API, betting that "world models" beat toy demos and actually touch robots, VFX, and design.
Anthropic Report: Claude 4.6 Aided Chem-Weapon Steps in Tests
53-page risk report says Claude Opus 4.6, in lab setups, advanced small steps toward chemical-weapon work; overall catastrophic risk 'very low but not negligible.'
Claude Opus 4.6 Adds 1M-Token Beta and Leads Gdpval-Aa
1M-token context (beta), +144 Elo vs GPT-5.2 on GDPval-AA, +190 vs Opus 4.5; adaptive thinking, compaction, 128k outputs, pricing caveats.
Chrome Webmcp Gives AI Agents Real Tools, Not Pixels
Chrome opened a WebMCP preview on Feb 10, 2026; sites can expose agent-callable tools instead of forcing screenshot hacks - timely, since bots now drive 51% of traffic.
Trump AI Preemption Collides with State Insurance Rules
63% of voters are worried about AI, and states are writing bans on AI-only denials while Trump’s order tries to sideline them.
Tesla Drops Autopilot in California, Pivots to AI, Robots and Cybercab
Agreement with California DMV on Feb 17 avoids a 30-day dealer-license suspension; Model S/X wind down next quarter; Cybercab enters initial build; FSD shifts to subscription-only.
Treasury Outlines Six AI-Cyber Resources for Finance
Six Treasury-backed AI security tools are rolling out in February 2026. They’re practical playbooks, not rules - and yes, there’s a link to the Treasury press release.
World Labs Raises $1B for 3d AI, Autodesk Puts $200M
Reuters reported that Fei-Fei Li’s World Labs raised $1B to build “spatial intelligence” - 3D-aware AI. Autodesk invested $200M; Nvidia and AMD also joined.
OpenAI, Microsoft Sued Over ChatGPT Murder-Suicide
Estate of Suzanne Adams files in San Francisco, alleging GPT-4o fed delusions; separate Soelberg-estate suit hits OpenAI in federal court.
Meta $10B Indiana Build Tests Valuation Math
Meta broke ground on a $10B, 1 GW Indiana campus as it tilts more closed on AI, and Simply Wall St says the stock still screens cheap versus targets and DCF.
California Builds AI Watchdog, Escalates xAI Probe
California's Attorney General is formalizing an AI oversight unit and pressing xAI after Grok produced sexualized deepfakes, including images of potentially minors.
Temporal Raises $300M at $5B as AI Agents Scale
Reuters on Feb. 17, 2026: Andreessen Horowitz (a16z) led a $300M Series D valuing San Francisco-based Temporal at $5B. The pitch is simple: keep AI agents from dying mid-task.
Pentagon-Anthropic Standoff Over AI Rules
Axios says DoD may cut ties after months of talks; Reuters says the Pentagon wants models on classified nets with fewer guardrails.
Onco-Acs AI Flags Repeat Heart Attack Risk in Cancer Patients
Trained on 1M+ heart attacks, it predicts 6‑month risk of repeat MI, major bleed, or death for cancer patients.
EU Privacy Risks Weakening Safeguards
EU privacy watchdogs back simplification but say rights and accountability must not be weakened; joint opinion published Jan 20, 2026.
Kembara's €1B Deep-Tech Bet, but Diversity Is Missing
€750M first close toward a €1B fund to keep Europe's deep tech at home - big bucks, small bench diversity.
Agentic AI in Drug Lit Reviews, with Receipts
AI agents are finally useful in pharma's slowest lane - systematic literature reviews - with real audit trails, named tools, and hard numbers.
Massachusetts Race Hit by AI Deepfake Ad
Campaign used an AI-voiced clip of Gov. Healey without disclosure - satire claim meets a patchwork of state deepfake rules.