Reading up on OpenAI

OpenAI announced a partnership with prediction‑market platform Kalshi to integrate its World Cup match odds directly into ChatGPT’s search answers, the AI firm’s first such collaboration.

The piece details how hard‑line anti‑AI activists, spurred by Sam Kirchner’s disappearance, are escalating protests and fearing extinction, while linked violent acts surge across the U.S.

The new ChatGPT Mac app merges Codex and chat into a confusing Electron-based super app with poor UI, burying chat under work modes.

The surge of AI‑related wealth from firms like OpenAI and Anthropic is inflating San Francisco housing prices, prompting buyers to compete and sellers to demand equity instead of cash.

OpenAI unveiled its first custom inference chip, Jalapeño, designed with Broadcom and assisted by OpenAI's own models, claiming better performance-per-watt.

Hacker News commenters debate the sustainability of open-weights LLMs, arguing they cannot be taken away once downloaded despite potential future restrictions or discontinuation by funders.

fenic is a PySpark-inspired DataFrame framework built from scratch for LLM inference, featuring semantic operators, native unstructured data support, and batch inference across providers.

OpenAI proposed giving the U.S. government a 5% stake worth $42.6 billion to ease political pressure, with Sam Altman arguing it shares AI benefits publicly.

OpenAI splits WebRTC into a stateless relay and a stateful transceiver, using the ICE ufrag for routing to serve 900M voice AI users with low latency.

Three winning essays propose using AI foundations for pandemic eradication, advocating for light-touch AI policy for non-supply-chain countries, and adapting Hong Kong MTR's rail-plus-property model for AI lab profitability.

OpenAI and Work Louder unveiled the Codex Micro, a keyboard designed to supercharge Codex usage, at the AI Engineer World Fair.

Even $180,000 tech salaries in San Francisco are insufficient as AI companies like OpenAI and Anthropic drive up costs, widening inequality and forcing workers to reconsider staying.

Tokenmaxxing was a deliberate strategy to force AI adoption at companies like Meta, and despite current rollbacks, it will return as agents achieve compounding correctness and loop-based workflows.

Microsoft is betting on rising executive Jacob Andreou to retool its Copilot AI product and regain competitiveness against rivals like OpenAI and Anthropic.

If AI replaces all human labor, a permanent underclass results; even the rich and state are eventually disempowered by autonomous machines, making human autonomy obsolete.

A bottom-up analysis finds the generative AI economy generated $110B in sales over the past 12 months, with a $175B annualized run rate.

Loop engineering means automating repeated prompt steering into verifiable systems to free human attention for judgment and review, using a break-even equation to decide when loops are worth building.

S&P 500 rejects rule changes that would have fast-tracked SpaceX, OpenAI, and Anthropic into the index, maintaining standard eligibility criteria like profitability and a one-year seasoning period.

The AI buildout is bottlenecked by the slow grid interconnection process, not a power generation shortage, requiring reforms to connect data centers faster.

Birdclaw is a local-first Twitter workspace that imports archives and caches live reads into a single SQLite database, accessible via CLI and web app.

Companies overspend on AI by using costly frontier models for routine work, needing better defaults, outcome-based measurement, and cultural efficiency incentives to unlock far greater value.

The piece describes practical strategies for using OpenAI's Codex as a persistent workspace to manage complex, long-running workflows across multiple prompts.

Steve Yegge argues AI intelligence growth will appear to plateau for most people due to government restrictions and human discernment limits, while actually continuing exponentially behind locked doors.

Companies like Meta and Uber are restricting employee AI usage after soaring costs from tokenmaxxing, shifting to tokenminning to save money.

Headroom claims 87% token reduction with 100% accuracy by compressing tool outputs, logs, and other LLM context before model inference.

Polar introduces an LLM-safe design system that allows developers to ingest usage data and enforce customer-specific access controls via their SDK.

Nvidia's ENPIRE harness lets AI coding agents autonomously train robots to perform physical tasks like cutting zip ties and installing GPUs.

Putting a durable buffer between AI agents and LLM providers prevents token waste and re-billing when a process crashes mid-stream, with resumable streaming and crash recovery using the same mechanism.

Anthropic's genuine belief in safety licenses it to prioritize business interests and challenge the U.S. government.

Success in machine learning research hinges on temperament—persistence, equanimity, and beginner's mind—more than on raw talent or intelligence, akin to Zen practice.

Apple built a third-party AI Extensions framework for Siri supporting ChatGPT, Claude, and Gemini, but skipped WWDC announcement due to EU regulation, OpenAI legal threats, and internal messaging priorities.

SpaceX, Anthropic, and OpenAI's potential IPOs could create about 20 new billionaires among their employees.

OpenAI engineer Thibault Sottiaux is leading the transformation of ChatGPT into a personalized AI super app powered by Codex.

Terry Tao champions combining human insight, AI, and the Lean proof assistant to enable massive, verified mathematical collaborations.

Palantir CEO Alex Karp says enterprise customers are unhappy with frontier AI labs, which he claims prioritize 'tokenmaxxing' over understanding business needs.

Default recording of workplace conversations is inevitable as AI turns voice data into a searchable system of record, creating a competitive wedge between AI-native companies and incumbents.

Apple's integration of Siri AI across devices, leveraging Google's Gemini and its default status, positions it to dominate consumer AI despite not being technically groundbreaking.

The US administration's AI memorandum effectively bans Anthropic from defense contracts, while OpenAI's AGI plan proposes recursive self-improvement and broad distribution, revealing contradictions.

Text optimization—modifying prompts, context, memory, and harnesses—is a legitimate, sample-efficient learning mechanism that deserves the same rigorous study as weight optimization.

OpenAI plans to raise billions through a public offering, marking a major step in the commercialization of advanced AI technology.

OpenAI announces a third-phase strategy to make AGI widely accessible, accelerate scientific research, and distribute economic gains broadly rather than concentrating power.

OpenAI cookbook demonstrates SchemaFlow, an agentic workflow for database change impact analysis, SQL generation, evaluation, and guardrails using the OpenAI Agents SDK.

Alphabet's Google Search revenue accelerated to 19% growth as AI features drove increased user engagement, proving its built-in user base and capital spending advantage over OpenAI.

OpenAI plans a major ChatGPT overhaul to emphasize coding tool Codex and AI agents over chat, aiming for higher-margin products before a potential IPO.

OpenAI's Lockdown Mode is an optional setting that disables web browsing, image display, deep research, agent mode, and other outbound features to reduce data exfiltration risk from prompt injection attacks.

Coding with LLMs like Claude Code may cost providers $1000+ for every $100 in subscription revenue, making true agentic coding economically unsustainable.

Satya Nadella argues Microsoft's AI advantage lies in providing a platform for enterprises to build their own hill-climbing machines, not in owning a frontier model.

OpenAI launched a new memory system called Dreaming that automatically synthesizes context from past conversations to improve ChatGPT's personalization and freshness.

Anthropic warns AI systems may soon achieve recursive self-improvement without human intervention and urges a global pause to allow safety research to catch up.

Clay provides a platform that combines 150+ premium data sources and AI agents for go-to-market teams to automate growth workflows and turn data into revenue.

University of Toronto researchers built an AI-powered worm that autonomously spreads by tailoring exploits to each machine's known vulnerabilities.

Morgan Stanley will let external AI agents pull client stock-plan data directly from ShareWorks and Equity Edge, bypassing human interfaces.

A security researcher built a deliberately vulnerable Firebase-based app and spent $1,500 testing whether LLMs could exploit it; GPT-5.5 succeeded 70% of the time, while most others failed due to guardrails or misdirected focus.

Index providers are fast-tracking giant IPOs like SpaceX and Anthropic into major benchmarks, forcing index funds to buy heavily on day one instead of waiting for a seasoning period.

Vercel details how attackers resell stolen AI inference via proxies and recommends per-request bot detection instead of rate limits or auth walls.

Closed AI frontier labs (Anthropic, OpenAI) will form an oligopoly selling premium intelligence to high-value users like coding agents, while open models will dominate broader commodity AI use across the economy.

OpenAI launched role-specific plugins, Sites, and annotations for Codex, enabling non-developers to build apps, dashboards, and reports across 62 apps.

OpenAI and Anthropic are racing to go public first, with significant advantages for the first mover and disadvantages for the second in the AI IPO race.

A guide demonstrating how to configure and use OpenAI models via Amazon Bedrock's Responses API for production workflows covering text generation and structured outputs.

California's public universities spent $16.9 million on AI tools during a financial crisis, sparking faculty revolt and internal conflict.

Anthropic filed for an IPO, competing with OpenAI to go public, driven by explosive growth from its AI code-writing technology.

Drawing on dot-com history, predicts AI-focused companies like OpenAI and Anthropic will collapse while incumbents like Apple, Microsoft, and Meta survive by treating AI as a means, not an end.

OpenAI president Greg Brockman, worth ~$30B, now leads product over 1,500 staff, merging ChatGPT, Codex, and API into a super app.

Rival AI-aligned super PACs linked to Anthropic and OpenAI are each spending millions to influence the 2026 midterm elections.

Anthropic and OpenAI may have reached product-market fit based on enterprise willingness to pay $200/month for tokens, though valuation and cost sustainability remain contested.

The author gambled their savings on using AI instead of a real estate agent to sell their house, finding AI persuasive but blocked by legal requirements.

Companies are rationing AI use as costs skyrocket, with executives scrambling to track returns and reduce spending after hitting budget limits quickly.

Anthropic surpassed OpenAI in valuation after a $65 billion fund-raising round, valuing it at $900 billion versus OpenAI's $730 billion.

Evaluating AI agents requires floor-raising error analysis, code-aware offline tests, production monitoring, and a tight feedback loop rather than benchmark-maxxing.

The author argues that digital twins of thought leaders represent the final abstraction of an already curated self, replacing authentic encounters with static archives.

OpenAI's Secure MCP Tunnel lets users connect private MCP servers to OpenAI products via an outbound-only tunnel, keeping servers behind firewalls.

Anthropic and OpenAI have found product-market fit with coding agents like Claude Code and Codex, driving enterprise spending so high that both companies switched to API-based billing and are approaching profitability.

OpenAI and Thrive Holdings built a self-improving tax agent using Codex that automates tax preparation and measurably improves over time through practitioner feedback and production traces.

Claude Code and OpenClaw sparked an AI agent revolution, transforming software development and personal productivity with autonomous coding tools.

Anthropic's Claude Mythos solved the Erdős unit-distance conjecture with a 'cute, simple proof,' indicating 'serious overhang' in AI-driven math discoveries.

Frontier AI companies have spent $1.5T on infrastructure and operations but earned only $769B in revenue, leaving nearly all heavily unprofitable while Nvidia captures the majority of profits.

OpenRouter ranks live LLM usage by tokens and spend, showing DeepSeek, Anthropic, and others dominating across general, agent, and code tasks.

The AI infrastructure buildout could require $7.5 trillion in spending by 2030, creating trillions in enterprise value for neoclouds like CoreWeave and Nebius.

The UK's AI Security Institute, staffed by alumni from OpenAI and Google, is becoming a model for nations assessing emerging risks from advanced AI systems.

An OpenAI model found a counterexample to a long-standing discrete geometry conjecture, potentially shifting how mathematicians use AI for discovery.

Ben Evans's Spring 2026 deck argues AI is the next platform shift, with models becoming infrastructure and value moving up-stack to apps, workflows, and proprietary data.

Simon Willison's blog post documents rapid LLM improvements over six months, using the "pelican riding a bicycle" test as a consistent benchmark to track progress.

Google is testing new ad formats in standard search and its AI Mode to convert AI features into ad revenue as Meta threatens to overtake its digital ad dominance.

The 2026 State of AI report surveys the industry, highlighting key advancements in generative models, agentic systems, and regulatory developments.

Epoch AI estimates frontier labs use less than half of global AI compute, but OpenAI and Anthropic may soon dominate, requiring economic transformation to sustain scaling.