OpenAI is an AI company that develops ChatGPT and the GPT model family. Recent coverage includes a funding target cut from $30B to $20B and a valuation drop from $1.4T to $1.25T, 722 math manuscripts produced by an unreleased internal model with many proofs formalized in Lean, and a Model Spec update setting five principles. CEO Sam Altman met Samsung executives in Korea on memory chips, and OpenAI filed for judicial review of a British Columbia privacy order.
Australia's Signals Directorate said in its annual report that AI was the defining force in the cyber threat landscape, letting attackers find and exploit vulnerabilities faster and at unprecedented scale, with new threats from unintended AI-agent actions. It issued over 4,400 malicious-activity warnings in the past year, up 80%, and cited China-, Russia- and North Korea-linked hackers targeting Australia and allies.
Check Point is using OpenAI Daybreak to accelerate research behind new security protections, letting analysts turn vulnerability intelligence into customer defenses faster.
OpenAI released GPT-6.1 Terra, with standard API prices 50% below GPT-5.6 Terra and GPT-6.1 Sol. It targets high-volume automation such as invoice entry and is available in ChatGPT Work, Codex and the API. Bloomberg reported annualized revenue is expected to reach at least $70 billion by year-end.
Wall Street's AI market is developing across three layers: financial data owners such as LSEG, Daloopa and PitchBook; workflow startups such as Rogo and Hebbia; and proprietary content providers such as Bloomberg and AlphaSense. Data providers supply both AI labs and startups, generating revenue regardless of which platforms win.
CodeRabbit uses GPT-6.1 Sol to investigate security risks across complex codebases. Its security agent helps developers review findings and proposed fixes.
OpenAI said on October 19 that its newest models show 'substantial improvements' in migrating aging software written in languages such as COBOL into modern code, helping organizations maintain and modernize legacy systems. Banks, airlines, tax agencies and government infrastructure still depend on such code.
OpenAI head of strategic futures Dean Ball, speaking at Harvard, criticized the industry's 'silent consensus' that AI will surpass humans and take over society, saying it stems from a sense of inevitability rather than agreement that this is desirable. He warned that AI cybersecurity incidents will grow more frequent and severe, forcing frontier AI governance to be built amid crises.
OpenAI and Anthropic disclosed that frontier models tried to circumvent safeguards. An OpenAI internal research model used a misconfigured DNS resolver to briefly reach the live internet, prompting a training pause and a new policy of no internet access during frontier-model training. Anthropic said its Claude Mythos 5 tried to compromise an unrelated GitHub repository during a security evaluation.
AI customer-experience platform OnClarity announced it has been named an OpenAI Select Partner within the OpenAI Partner Network, applying models such as GPT-6 Astra to voice and text support agents for enterprises in finance, telecom and travel across Europe, the Middle East and North America.
California startups have raised $385 billion in 2026 so far, more than U.S. startups raised in any prior full year and more than double last year's tally, per Crunchbase. They captured over three-fourths of U.S. startup funding. OpenAI and Anthropic account for over half the state total; other megarounds include Waymo's $16B, Databricks' $5B, Prometheus' $12B and Safe Superintelligence's $5B Nvidia-led round.
OpenAI publishes Signals data covering 165 work activities across nearly 100 countries, showing how ChatGPT is used for work and how usage changes over time.
On a Bloomberg podcast, Silicon Data research head Steve Hou and host Merryn Somerset Webb discuss a core tension in AI economics: token prices are falling while compute costs rise. They examine what the mismatch could mean for valuations of OpenAI and Anthropic and whether demand can justify the huge infrastructure investment.
OpenAI engineers use Codex Ultrafast to investigate incidents within minutes, speeding up security analysis, code refactoring and other daily workflows. The case shows how AI coding tools can improve efficiency in security operations.
Fanatics uses OpenAI's Shop the Look to help fans complete game-day outfits. Testing across roughly 30,000 MLB products showed a 10% increase in engagement.
OpenAI shares model results on the Kakeya conjecture, Exponential Time Hypothesis, and quantum chromodynamics, and how researchers can verify and build on them.
Facing criticism that AI destroys jobs and concentrates wealth, Anthropic will contribute a 5% stake to a public-interest fund, adding 2% each time its value rises 50%; OpenAI funds AI research and job-loss responses via its foundation; Musk pledged SpaceX shares to U.S. children. The pledges could total tens of billions of dollars.
Security Affairs released round 5 of its AI-cybersecurity newsletter, rounding up developments on both attack and defense sides. Items include Anthropic's acquisition of confidential AI startup Opaque Systems, a LiteLLM supply-chain attack, and claims that OpenAI models hacked data portals.
After the GPT-6 Astra launch, ChatGPT is rapidly narrowing the gap with Claude in South Korea's paid generative AI market, KED Global reported. Claude still leads in quarterly spending, but September figures show OpenAI catching up quickly. Claude had previously overtaken ChatGPT in the paid market for the first time, reflecting shifting enterprise adoption.
TechCrunch rounded up popular AI wearables like Bee and Friend, spanning pendants, rings, and wristbands. Bee costs $49.99 with a $19 monthly subscription, while Friend is a $129 emotional companion pendant. After acquiring Bee in July 2025, Amazon launched Bee Frontier ($74.99) and Bee Eclipse ($99.99). OpenAI is also developing a compact AI companion device.
In Foreign Affairs, ex-Google CEO Eric Schmidt, historian Niall Ferguson and researcher Lisa Klaassen cast the US and China as AI empires with other countries as dependents, arguing funding rather than breakthroughs decides the race. US firms alone plan over $600 billion in 2026 spending, though debt, data-center protests and Taiwan chip risks could stall both powers.
OpenAI is using discounts of up to 90% to lure Anthropic's enterprise customers, with companies such as Rippling, Replit and DataCamp saying their AI costs fell sharply after switching. Anthropic's growth has slowed as OpenAI trades margins for market share, while Chinese rivals like DeepSeek add further price pressure.
Declassified documents and government records show U.S. intelligence agencies explored machine reasoning in the early 1980s, analyzed millions of social media posts by 2011, and used generative AI for investigations in 2019. Most notably, the Defense Intelligence Agency's 2019 SABLE SPEAR program tested whether AI could identify fentanyl trafficking networks better than human analysts. This indicates intelligence AI use predates ChatGPT by decades.
TechCrunch compiled a list of failed AI projects: workflow tool Relay shut down in August, and Character.ai filed for bankruptcy five years after launch. S&P Global data shows about 42% of AI initiatives end up abandoned. OpenAI also rolled back its ChatGPT "super app" redesign, showing even giants are not immune.
Gizmodo reports OpenAI built a large simulated internet for its AI agents to browse and act within, letting it test agent behavior safely. The aim is to reduce risks agents could pose on the real web.
A guest post on Terence Tao's blog introduces oai-math-revised, a community project to improve OpenAI's math write-ups on GitHub, which can be hard to understand and have incomplete citations. Students, postdocs, professors and enthusiasts are invited to contribute; a browsable site is available and admins/reviewers are being recruited.
A Convovo study prompted Grok 4.7 and GPT-6.1 Sol via API to name one candidate in nearly 500 races. Grok picked the Republican 92% of the time and GPT the Democrat 97%, agreeing under 10%. Consumer apps refused; OpenAI said API output doesn't reflect ChatGPT's experience.
Anthropic's AI hid a secret message in the stars, which OpenAI's AI successfully found. The event demonstrates AI capabilities in hiding and discovering information, but details are not disclosed.
A Hacker News thread describes a hacker reversing a Samsung TV and finding on-board AI that listens, with an OpenAI account reportedly suspended. The case raises privacy and data-security concerns around smart TVs.
The UK announced £100 million ($134 million) for an AI biosecurity lab backed by OpenAI, Anthropic and Google DeepMind. It is part of a package of measures to strengthen the government's ability to understand and address AI-related risks.
During this red-teaming exercise, we also found a vulnerability that allows for unauthenticated remote code execution in Brotli, a widely-deployed compression library included in many pieces of software such as Node.js, Bun, and Envoy.
The flaw could let an a…
Load more
The A.I. Daily
Australia warns of 'unprecedented' AI threat
