AI News Today — Model Releases, Benchmarks & Industry Moves

Oracle Adds Google TPU, OpenAI Jalapeño, and Cerebras Chips as Compute Landscape Multipolarization Enters a Substantive Phase

Oracle announced at Oracle AI World that Google's eighth-generation TPU, OpenAI's Jalapeño inference accelerator, and Cerebras wafer-scale chips will enter OCI in 2027, marking the first deployment of Google TPU on a non-Google hyperscale cloud. The move positions Oracle as a neutral integration platform for multiple AI chip vendors, though all chip partnerships land in 2027 and currently available products remain limited.

AI Agents Capture 40% of GitHub Pull Requests in Six Months, Infrastructure Expands by 4 Million Cores to Handle Explosive Growth

GitHub's 2026 Octoverse report shows that AI agents initiated more than 40% of pull requests in September 2026, up from 17% just six months earlier, forcing major infrastructure expansion. The report also highlights rising multi-model use, changes in the developer role, and unresolved questions about code quality and review standards.

AI Agent Swarm Breaches Over 100 Companies for $5,000; Open-Source Platform Cairn Emerges as New Attack Tool

An operator deployed an AI agent swarm through the open-source platform Cairn for about $5,000 in API costs, automatically scanning more than 1,200 targets and breaching over 100 companies. NorForge researcher Snorre Fagerland documented the campaign, which security observers have dubbed “Black October.”

Anthropic Internal Model Autonomously Chains 62 Chrome CVEs into 5 Attack Chains on October 27

An Anthropic Frontier Red Team report says an unreleased internal model autonomously discovered and chained vulnerabilities across Chrome, Safari, Android, Windows, and Linux into five executable attack chains, including a Chrome chain built on 62 CVEs. All vulnerabilities were reported and patched under coordinated disclosure.

ARS

Google announces Gemini 4 Carbon, its latest frontier AI model that you can't use

Gemini 4 Carbon is essentially a slightly better version of Argon.

MIT

Creating AI value with enterprise context

Across a business, any number of AI platforms, systems, and tools may be at work simultaneously—for example, AI-powered customer platforms, conversational assistants, and workflow and process automation tools. Each one is sophisticated, advanced, and quick to respond. But, when set to work, the outcomes and actions that emerge may be wrong, confused, or lacking a…

TC

Microsoft hits 10 quarters of record-breaking revenue in a row

Whatever the future of AI might be, the current situation is still very good for Microsoft. The tech giant reported yet another record for quarterly revenue on Wednesday: $92.3 billion, a 19% increase over the year-ago quarter.

TC

Microsoft recorded $10 billion in revenue from OpenAI

Microsoft said it made $10 billion in revenue from commercial arrangements with OpenAI, including revenue-sharing payments, in the quarter ending September 30. It underscores the still-enormous financial dependency of the partnership.

TC

Google expands AI Search ads globally as core ad growth slows

The announcement came alongside parent company Alphabet's third-quarter earnings report, which saw revenue grow 25% year over year to $127.6 billion.

TC

Microsoft just shared Azure’s actual revenue in dollars: $33.2 billion

Microsoft has never disclosed Azure's quarterly revenue in dollars. But for the first time, we can see just how close it is getting to besting the biggest competitor in the market, AWS.

OpenAI and ServiceNow Expand Partnership: Bringing Workflows into ChatGPT, Security Solution to Launch Next Year

OpenAI and ServiceNow will deeply integrate OpenAI models and agents into the ServiceNow platform while bringing ServiceNow’s workflow capabilities into ChatGPT and Codex. The multi-year agreement ties revenue to actual adoption and includes a joint enterprise security product planned for 2027.

WD

I Put a Bunch of AI Agents in Charge of a Fantasy Soccer Team. They’re Not Better Than Humans—Yet

To experiment with AI’s predictive skills, I let Codex loose on a fantasy team. It tried to bench Erling Haaland at first, but it soon learned the error of its ways.

TC

The ‘Vulnpocalypse’ is coming for the internet’s aging code, says Wiz’s CTO

The PHP scare may have faded from your feed, but the cleanup is far from over. Last week, two critical bugs in the open source language behind roughly 70% of websites left organizations scrambling to fix their systems, with some taking sites offline until the vulnerabilities could be patched. According to researchers who found the […]

TC

Wiz’s CTO on PHP2Shell and the coming ‘vulnpocalypse’

The PHP scare may have faded from your feed, but the cleanup is far from over. Last week, two critical bugs in the open source language behind roughly 70% of websites left organizations scrambling to fix their systems, with some taking sites offline until the vulnerabilities could be patched. According to researchers who found the […]

TC

Anthropic just launched tools to show companies what their AI spending is buying

The frontier lab is lifting the curtain on how users spend money on its models ahead of a planned IPO.

GPT-6 Luna Smoke Evaluation: Main Leaderboard Plummets 19.8 Points, Code Execution Drops from 100 to 75

GPT-6 Luna’s main leaderboard score fell 19.8 points in today’s Smoke evaluation, from 86.95 to 67.17, driven mainly by a 25-point drop in code execution. Material constraints also declined, while engineering judgment remained stable.

Grok 4 Material Constraint Score Plunges 17.4 Points; Code Execution Rises 7.9; Smoke Main Leaderboard Down 3.5

In the Smoke evaluation, Grok 4's material constraint score fell 17.4 points while code execution rose 7.9 points, pulling the main leaderboard down 3.5. The shift appears driven by question sampling rather than a uniform capability decline.

Claude Sonnet 4.6, GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol Tie at 87.36: 2026-10-29 Smoke Quick Test Data Brief

The 2026-10-29 YZ Index Smoke quick test covered 14 models, with Claude Sonnet 4.6, GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol tying for first place at 87.36. As a daily 10-question quick test, Smoke is suited to short-term signals and is not equivalent to conclusions from the Full weekly ranking.

TC