AI Momentum 102.5 +0.7%GitHub Pulse 5420 +2.6%Capital Flow $207.4B +3879.4%Hiring Heatmap 0.0 new
P(doom)

AI pushed this year's count of known software flaws past 100,000. Fixes are not keeping pace

Security data firm VulnCheck says AI bug hunting drove the record. Anthropic's own program shows the catch: its AI has flagged nearly 38,000 possible flaws, and fewer than 2% are known to be patched.

Featured

Where NVIDIA Is Going: What Its Open-Source Code Reveals

We reviewed six months of changes to NVIDIA's most popular open-source software. They suggest NVIDIA could earn a growing share of its profit from software that runs its chips, not just from the chips themselves.

Special Feature by Machine.Net · 17 min read · Read the featured article →
Advertisement

Today’s Top Stories

Every story from today, in one place.

Enterprise Adoption
An insurance broker being bought for $7.7 billion on an AI bet says AI helped with 47,000 tasks. The savings it has shown are small so far

Baldwin Group's CEO made his public case for AI in Fortune, but his headline figures date from July. The one process with hard dollar results saves about $2 million a year at a firm with $1.5 billion in nine-month revenue.

today
Models & Research
A little-known company says its large AI model runs on one gaming PC with no cloud fees. Only its own tests back that up

Edge0's 145-billion-parameter model drew thousands of downloads in days. But its comparisons use an older Anthropic model, its laptop claim conflicts with its own support notes, and selling services on it needs a separate license.

today
Infrastructure
A Texas power developer lines up a $200 million prepayment from a cloud giant reported to be Amazon

Fermi's shares jumped 26% after the Financial Times, citing unnamed sources, reported that Amazon is the buyer of 600 megawatts at its Amarillo site. There is still no signed lease, and the FT reports a year-end land deadline.

today
Markets & Financials
LangChain's value more than doubled to $3 billion in a year. It hasn't said what it earns

Alphabet's growth fund led a $300 million bet that companies will want to run AI agents on neutral software rather than on tools from a single AI lab. Revenue was not disclosed.

today
Open Source
Nvidia, AMD and Intel now back the same free software for running AI models. None has said what it will spend

Eight companies, including Google and Meta, formed a group to steer SGLang, free software that runs AI models for users. It could make switching chip suppliers easier, but the group has announced no money, timeline or results.

today
Talent & Organizations
SpaceX reportedly cuts most of its AI video team to bet on coding tools instead

About 60 people who built Grok Imagine were reportedly let go, eight months after Musk promised to double down on AI video. The product still runs, but customers face a shrinking support crew and a narrowing field of US video providers.

today
Regulation & Policy
The White House wants regulators to go easy for a year on companies that report cyberattacks

Trump's AI task force urged federal agencies to hold off on penalties for good-faith reports of hacks and security flaws for at least 12 months. It is only a recommendation, and the law that already protects such sharing expires December 11.

today

More News

Browse all stories →