|
📖 Read In Depth
|
Sam Altman may control our future – can he be trusted?
Ronan Farrow and Andrew Marantz's 18-month investigation into Sam Altman — alleges a pattern of deception, billions in funding from Gulf states, a cover-up with no written report, and reveals Elon Musk has been running a surveillance operation against him. The author himself showed up in the HN comments to field questions. This is the kind of long-form accountability journalism that actually matters when one person controls the most consequential technology company in the world.
hn/Best Stories
|
Issue: Claude Code is unusable for complex engineering tasks with Feb updates
A widely-upvoted GitHub issue documenting that Claude Code regressed significantly on complex engineering tasks after a February update — with detailed breakdowns of specific failure modes. A Claude Code team member (Boris) responded directly in the thread. If you're using Claude Code for real engineering work, this is actionable signal about what broke and what Anthropic's response is.
hn/Best Stories
|
The cult of vibe coding is dogfooding run amok
Bram Cohen (BitTorrent creator) argues that vibe coding — letting AI generate code without understanding it — is dogfooding run amok, producing systems nobody can maintain or reason about. The HN thread is equally worth scanning: there's a sharp counter-argument that Claude Code itself was allegedly built this way and still shipped a wildly successful product, which cuts right to the heart of whether craft and comprehension still matter at scale.
hn/Best Stories
|
A cryptography engineer's perspective on quantum computing timelines
A cryptography engineer's sober, detailed take on when cryptographically-relevant quantum computers (CRQCs) will actually arrive — pushing back on both the alarmists and the dismissers. Relevant for anyone thinking about post-quantum migration timelines in real systems, and a good model for how to reason carefully about uncertain technical futures.
hn/Best Stories
|
Axios: Sam Altman States Superintelligence Is So Close That America Needs A New Social Contract On The Scale Of The New Deal During The Great Depression
Sam Altman is publicly claiming superintelligence is close enough that the U.S. needs a New Deal-scale social contract overhaul — wealth funds, 4-day workweeks, redistribution mechanisms. This isn't just hype; OpenAI published an actual policy blueprint. Whether you think the timeline is credible or not, the fact that the most powerful AI lab is actively lobbying for these economic structures matters.
reddit/r/singularity
|
The Big Bang: A.I. Has Created a Code Overload
AI-generated code is now shipping faster than human engineers can review, test, or understand it — creating a 'code overload' problem at companies across Silicon Valley. Directly relevant to Xinyu's interest in the craft of software engineering and what it means when the production pipeline floods with AI output nobody fully owns.
nyt/Technology
|
A.I. Is on Its Way to Upending Cybersecurity
New AI systems from Anthropic and OpenAI are dramatically lowering the skill floor for offensive cyberattacks while also enabling faster defensive response — but offense may be winning. Covers real deployments by both red teams and actual threat actors, with specific examples of what's changed in the past year.
nyt/Technology
|
A New Oil Shock Accelerates a Return to Nuclear Power
The Iran war and resulting natural gas supply shocks are accelerating nuclear energy reconsideration in Asia and Europe — countries that swore off nuclear after Fukushima are quietly reversing course. A good piece on how geopolitical shocks reshape long-term energy infrastructure decisions, with real implications for data center power planning.
nyt/Business
|
Rival Nations Seize On Choke Points to Counter Trump
Iran and China are actively exploiting geographic and economic choke points — the Strait of Hormuz, semiconductor supply chains, rare earth exports — as leverage against U.S. aggression. A good structural piece on how adversaries are thinking about asymmetric pressure, with implications for chip supply chains and tech hardware costs.
nyt/Business
|
|
🎬 Check It Out
|
The multiple messages and themes in TAR (2022)
A substantive Reddit thread unpacking the layered themes in Tár (2022) — power, institutional rot, cancel culture, and the ethics of separating art from artist. If you haven't seen Tár, it's available on Peacock and is one of the most intellectually dense films of the decade. The thread is worth reading after watching.
reddit/r/TrueFilm
|
|
⚡ FYI
|
Anthropic expands partnership with Google and Broadcom for next-gen compute
Anthropic has signed a deal with Google and Broadcom for multiple gigawatts of next-gen TPU compute, with ARR now reportedly at $30B (up from $9B at end of 2025). This is a major infrastructure commitment that signals Anthropic is betting heavily on custom silicon rather than just NVIDIA GPUs — relevant for anyone tracking the compute layer of the AI stack.
hn/Best Stories
|
Ronan Farrow’s investigation into Sam Altman— alleges pattern of lying, billions from Gulf dictators, a cover-up investigation with no written report, and reveals Elon Musk has been running a full surveillance operation against him including hiring people to investigate his sex life at gay bars
Reddit thread summarizing key claims from the Farrow/Marantz New Yorker investigation, including the Musk surveillance angle. Skip this if you're reading the primary article above — but useful for the community reaction and additional context threads linked within.
reddit/r/singularity
|
Tech employment just fell below 2016 levels. Here is where capital appears to be rotating
US tech employment has dropped below pre-pandemic (and even 2016) levels while AI investment surges — a stark data point on where the labor displacement is already showing up. The post maps where capital appears to be rotating instead, which is useful context for thinking about what the AI productivity wave is actually doing to the sector.
reddit/r/investing
|
Can Science Predict When a Study Won’t Hold Up?
A major study finds that AI is not yet reliable for predicting whether scientific studies will replicate — it can pattern-match surface features but doesn't capture the deeper methodological signals that humans use. Relevant pushback against the 'AI scientist' hype, with implications for anyone thinking about AI in research workflows.
nyt/Technology
|
Justice Dept. Resolves Broadway Touring Company Investigation
The DOJ resolved its investigation into Broadway Across America after the touring company acknowledged signing anticompetitive noncompete agreements with rival presenters — declining to prosecute but extracting an acknowledgment. Part of the broader Live Nation antitrust story and how entertainment conglomerates use contracts to foreclose competition at the regional level.
nyt/Business
|