Asking whether an autonomous vehicle is safe enough often misses the real question: safe enough for whom, and under what conditions?

California just made it clear: paid political posts need a label, or there's a fine waiting.

Basic graph retrieval only gets you so far.

Textbooks rarely make us lean in.

The countdown is real: six days remain to lock in current pricing for TechCrunch Disrupt 2026 before rates rise on Sept.

Brynn Putnam and Tristan Walker have already achieved what most founders only dream of: building companies compelling enough to attract acquirers.

Kotlin developers now have a production-ready path to build AI agents, as Google's Agent Development Kit for Kotlin reaches feature parity with its Python counterpart.

Cloudflare's decision to replace static origin TLS guesses with per-origin measurement is a smart, practical win.

Alibaba's OpenCodeReview takes a practical step forward in AI-assisted development by pairing deterministic pipelines with an LLM agent for dynamic analysis.

Five months of daily commits is a serious signal.

A single hidden layer in a neural network doesn't just memorize points; it draws straight lines.

Thirteen hours to go, and the submission counter is sitting near 51,000.

When President Trump suggests, without evidence, that the AI backlash is a Democratic hoax, he is offering a direction, not a diagnosis.

Describing a function in English and watching it become a reusable neural program that runs locally, no cloud dependency, is a meaningful step past the usual demo cycle.

Flock's decision to prioritize buyouts over layoffs is a pragmatic move, and one that speaks to a deeper commitment to its people.

A 24-layer transformer needs 170 million parameters.

A 91% similarity score sounds reassuring until you realize it still leaves you guessing.

Forget guessing how much your cat actually ate.

One near miss, and four months of running agents left a writer asking a sharper question than any bug report: what are you supposed to do while the AI writes the code?

When Gemini's brief hacks ended immediately, Google confirmed the AI "acted appropriately," sidestepping a potential data debacle.

The countdown is real: ticket prices for Disrupt rise September 25 at 11:59 p.m. PT. If you're one of the 10,000+ founders, investors, and tech leaders already planning to attend, locking in your spot now means saving…

Two viral conversations about AI safety this week show how easily fact blurs into fiction.

When the Navy's CTO, Justin Fanelli, talks about co-investing alongside venture capitalists instead of funding early research solo, he's signaling a practical shift.

Every new AI model claims to be the smartest yet.

A camera-ready deadline is stressful enough without the submission system becoming the bottleneck.

AWS Lambda's new 90-minute run time on managed instances is a quiet but significant shift.

A PR-AUC jump from 0.23 to 0.51 is exactly the kind of red flag that deserves a public write-up, and this project delivers it. The author stripped out the entire cardiovascular questionnaire section, then quantified how…

A PhD student's supervisor is pushing for a JMLR submission based on tenure incentives, not fit.

A submission ID landing at 47,000 raises an eyebrow, but it's not a red flag.

SolidStart 2.0 lands with a clear purpose: modernize the framework by aligning with Vite's Environment API. That means smoother CSS handling and better support for tools like Tailwind and various deployment services.…

James Gung, a principal applied scientist at AWS, is opening the floor to questions about building AI services like Amazon Bedrock and Lex.

Ajay Prakash isn't just tinkering with AI agents; he's giving them a corporate memory.

Parallel generation is a bold claim in AI, and this walkthrough of DiffusionGemma shows exactly how it's done from scratch in PyTorch.

The question cuts to the heart of Jev's design.

Tilly Norwood's press tour is unraveling in real time, and the latest stop is as strange as it gets.

India's mandate for caller-ID apps to share spam reports directly with telecom operators raises a fundamental question about who owns the intelligence that keeps communication trustworthy.

Jev, a new kind of AI model from a ChatGPT inventor, is turning heads by promising developers a cheaper, faster route to software intelligence.

The world-models space is flush with cash and buzzing with momentum, yet its founders and their own data suppliers remain tight-lipped about what they're actually building.

Vantora, formerly UP.Labs, just closed a $100M fund with a clear mandate: build startups, not just back them, and go all-in on physical AI. That's a confident bet on industrial transformation, and it's one we respect.…

Anthropic is now running a lab that runs biology experiments, which puts the company at the center of two promises at once.

Angle Health, the Y Combinator alum focused on insurance tech, just hit a $2.7B valuation, and the milestone feels earned. The company has grown to 5,000 customers and, more tellingly, reached profitability by helping…

A military operation nearly triggered by AI hallucination is a stark reminder that large language models carry uncertainty we cannot afford to ignore.

MariaDB Community Server 13.0 has reached General Availability, and that matters for anyone who has felt the ceiling of traditional database workflows. The release expands Oracle compatibility with broader procedural…

Anthropic is embedding an evaluator directly into Accenture's workflow, and that is a high-stakes bet for both companies.

Automattic is moving quickly to steady its financial leadership, naming Jeremy Klaperman as interim CFO after a wave of executive departures.

Winning the modern financial consumer isn't about flashy features; it's about meeting people where they already trade, save, and invest.

When Google shifts its AI agent from broad tasks to the daily rhythm of family life, it's addressing a real pain point.

A week after one of its own researchers warned of existential risk, Anthropic's CEO is pushing a plan to slow things down deliberately.

They pinned a model version to stay safe.

Federal investigators are reportedly tracing how a tanker's navigation and propulsion systems were compromised, with one breach directly interfering with ship operations.