AI News Archive - Browse Page 89 of 182
Browse AI news articles covering LLMs, tools, research, and industry trends
Anthropic adds Claude Code Review, USD 20 per bug, sues Pentagon blacklist
Twenty dollars per bug. That’s Anthropic’s price tag for catching a flaw before it hits production, and the company knows it sounds absurdly low.
Google Stax uses LLM-as-judge to auto‑evaluate model outputs by your criteria
Evaluating AI outputs at scale is a bottleneck, one that Google Stax aims to dissolve.
Microsoft unveils Copilot Cowork, Anthropic AI that syncs M365 work
Microsoft’s most ambitious bet on AI just got a new brain, and it’s not built in Redmond.
Anthropic Files Lawsuit Against DoD Over Supply-Chain Risk Designation
The Pentagon can secure its supply chain. What it allegedly cannot do is rig the bidding.
LangChain GTM Agent pulls Salesforce, BigQuery data, tracks funding, launches, AI
Monday mornings, the agent checks in. It pulls Salesforce into focus, queries BigQuery for patterns, then scans the open web for funding rounds,...
Falling costs drive expansive accessibility to language models
The price of raw AI has plunged. Remember paying tens of dollars to process a million tokens with a top model? That bill is now pennies.
Microsoft says ungoverned AI agents could become double agents; USD 99/mo fix
The pitch is seductive: hire an AI agent for the price of a premium streaming subscription, $99 a month.
AI Service Providers Earn via Setup Fees and Ongoing Monthly Retainers
The profit model is brutally simple: a setup fee to build the workflow from scratch, then a monthly retainer to keep it alive. Support. Monitoring.
Survey: 1 in 5 adults misinterpret shared decisionmaking on vaccines
A single survey finding has exposed a dangerous fault line in public health communication: one in five adults now believes that "shared...
OpenAI robotics lead departs amid Pentagon contract, partners with You.com
The exit of OpenAI’s robotics lead wasn’t just a resignation, it was a signal. The departure came as the company deepened its ties with the Pentagon,...
AI Agents Poised to Rival Top Venture Investors, Threaten Sand Hill Road
Sand Hill Road has long been the beating heart of venture capital, a corridor of power where billions are deployed and unicorns are born.
OpenAI launches GPT-5.4 in standard, Pro, and Thinking versions
OpenAI is done pretending one model fits all. Its new GPT-5.4 comes in three distinct versions, and the choice you make says more about your work...
A2UI lets agents generate UI screens from a flexible UX schema
The next frontier of AI isn’t just about smarter models, it’s about how those models touch the screen. Enter A2UI: agent-to-user interface.
Rajan says AI-native fundamentals let new grads outpace senior devs
Experience is overrated. At least, that's the argument now. A new graduate who learns to think alongside AI can leave a senior developer, one who...
OpenClaw Superfan Meetup Highlights Optimism, Lobster and Varied Interests
The scent of melted butter and the hum of unfiltered ambition filled the room. At the OpenClaw Superfan Meetup, the vibe was unmistakably optimistic,...
Large CUDA Tiles Reduce Flash Attention TFLOPS by 18‑43% Across Sequences
Bigger tiles should mean fewer memory accesses, faster attention, right? Not on NVIDIA GPUs.
Anthropic releases Claude Marketplace for tools from Replit, GitLab, Harvey
Anthropic just made a power move that could reshape how enterprises buy AI. Claude Marketplace is here, and it flips an emerging narrative on its...
Jammer Aims to Block Always-Listening Wearables, Drops Flock Safety Deal
The pitch is seductive: a device that uses AI to fight AI, a jammer that silences the always-listening microphones embedded in our glasses, our pins,...
LangChain CEO says model quality alone won’t deliver production AI agents
Silicon Valley’s obsession with bigger, better models misses the real bottleneck.
KV cache compaction cuts LLM memory 50×, chunked processing long contexts
A research team slashed the memory load of large language models by 98 percent. Their secret? Compressing the model's key-value cache fiftyfold.