Older Coverage
stories 781–815 of 815
AI
Researchers Unveil 'Self-Harness,' Letting AI Agents Rewrite Their Own Rules for a 60% Boost
Researchers introduced Self-Harness, a framework that lets AI agents rewrite the scaffolding and rules that govern their own behavior, reporting performance gains of up to 60% on agentic tasks. The work pushes agents from static, human-authored harnesses toward systems that adapt their own operating instructions on the fly.
AI Hit the Memory Wall -- and Now It Needs a Whole New Context Tier
AI
AI Hit the Memory Wall -- and Now It Needs a Whole New Context Tier
VentureBeat argues that AI's biggest emerging bottleneck isn't raw compute but memory -- the gap between fast, expensive on-chip memory and the vast context modern AI workloads demand. The proposed fix is a new 'context tier' in the memory hierarchy purpose-built for long-context and agentic AI.
Crunchbase: Classic SaaS Isn't Coming Back -- Agentic AI Is Replacing It
AI
Crunchbase: Classic SaaS Isn't Coming Back -- Agentic AI Is Replacing It
A Crunchbase analysis argues the traditional SaaS model isn't recovering from its multi-year slump -- something bigger is supplanting it: agentic AI that does the work rather than just organizing it. The piece reframes the software downturn not as a cyclical dip but as a structural handoff from seat-based tools to outcome-based AI agents.
xAI's Grok 4.3 Goes Generally Available on Amazon Bedrock -- the Cheapest US Frontier Model There
AI
xAI's Grok 4.3 Goes Generally Available on Amazon Bedrock -- the Cheapest US Frontier Model There
Amazon made xAI's Grok 4.3 generally available on Bedrock, making xAI the third major independent lab on the platform alongside Anthropic and OpenAI. Priced at $1.25/$2.50 per million input/output tokens with a 1-million-token context window, Grok 4.3 lands as the cheapest US-lab frontier reasoning model on Bedrock.
Beyond Siri: The Practical AI Features Coming to Your iPhone in iOS 27
AI
Beyond Siri: The Practical AI Features Coming to Your iPhone in iOS 27
Apple is previewing a slate of practical, on-device AI features in iOS 27 that go well beyond its long-delayed Siri overhaul, per TechCrunch. The emphasis on concrete, everyday utility -- rather than a single flashy assistant -- signals Apple's bid to make AI feel useful and trustworthy to a billion-plus users.
'In the Weights' Lets You Search What AI Models Actually Know About You
AI
'In the Weights' Lets You Search What AI Models Actually Know About You
A new tool called In the Weights functions as an AI-centric vanity search, letting people probe what large language models 'know' or output about them, per TechCrunch. As models increasingly mediate reputation and information, tools that surface what they encode about individuals are becoming a new layer of the identity stack.
Signal's Meredith Whittaker: AI Chatbots 'Are Not Your Friends'
AI
Signal's Meredith Whittaker: AI Chatbots 'Are Not Your Friends'
Signal president Meredith Whittaker is warning users not to mistake AI chatbots for confidants, arguing the intimacy they simulate masks a surveillance-and-data business underneath. Her caution lands amid a boom in companion and assistant AI that increasingly positions itself as an emotional relationship.
Huawei-Led Team Post-Trains DeepSeek's 1.6-Trillion-Parameter Model on Ascend Chips, Not Nvidia
AI
Huawei-Led Team Post-Trains DeepSeek's 1.6-Trillion-Parameter Model on Ascend Chips, Not Nvidia
A research group including Huawei completed full-parameter post-training of DeepSeek's 1.6-trillion-parameter V4-Pro model using a cluster of at least 1,000 Huawei Ascend 910C chips -- demonstrating that a frontier-scale Chinese model can be trained end-to-end on domestic silicon. It's a milestone in China's drive to break free of US chips and the CUDA ecosystem.
Nobel Laureate John Jumper Leaves Google DeepMind for Anthropic, Deepening Google's AI Talent Drain
AINobel laureate hire
Nobel Laureate John Jumper Leaves Google DeepMind for Anthropic, Deepening Google's AI Talent Drain
John Jumper -- the 2024 Nobel laureate who co-created AlphaFold -- is leaving Google DeepMind after nearly nine years to join rival Anthropic. The exit lands just days after Gemini co-lead Noam Shazeer departed for OpenAI, turning a brutal week of poaching into the clearest sign yet that the AI race is now a talent war, and that Google's crown-jewel lab is the one bleeding.
Signal's Meredith Whittaker Warns AI Chatbots 'Are Not Your Friends'
AI
Signal's Meredith Whittaker Warns AI Chatbots 'Are Not Your Friends'
Signal president Meredith Whittaker is pushing back hard on the rise of companion-style AI, warning that chatbots designed to feel like friends are surveillance-driven products optimized for engagement, not your interests. Her argument lands as companion AI goes mainstream and regulators start to circle.
7,000 Langflow Servers Are Under Attack -- and LangGraph and LangChain Share the Holes
AI
7,000 Langflow Servers Are Under Attack -- and LangGraph and LangChain Share the Holes
Roughly 7,000 internet-exposed Langflow servers are being actively exploited, and researchers warn the same class of flaws extends to LangGraph and LangChain -- the orchestration backbone of much of today's agent stack. It's a stark reminder that the rush to ship AI agents has outpaced the security hardening underneath them.
7,000 Langflow Servers Are Under Active Attack -- and LangGraph and LangChain Share the Holes
AI
7,000 Langflow Servers Are Under Active Attack -- and LangGraph and LangChain Share the Holes
Roughly 7,000 internet-exposed Langflow servers are under active attack, and security researchers warn that the same class of vulnerabilities extends to LangGraph and LangChain -- the de facto plumbing of the agent boom. As enterprises rush AI agents into production, the orchestration layer is emerging as a soft, high-value target.
Hypernetworks Build the Exact Model Your Agent Needs, On Demand -- Where Fine-Tuning and RAG Fall Short
AI
Hypernetworks Build the Exact Model Your Agent Needs, On Demand -- Where Fine-Tuning and RAG Fall Short
A new approach argues that fine-tuning forgets and RAG leaks context, and that hypernetworks -- models that generate the weights of another model on demand -- can produce a task-specific model for an agent in the moment. It's a fresh take on the persistent problem of giving agents durable, reliable, situation-specific knowledge.
Microsoft Uncovers a New Self-Propagating Backdoor Built to Steal Cryptocurrency
AI
Microsoft Uncovers a New Self-Propagating Backdoor Built to Steal Cryptocurrency
Microsoft researchers discovered a new lightweight, self-propagating backdoor designed to drain cryptocurrency from infected machines. Its ability to spread on its own marks an escalation in financially motivated malware at the exact moment crypto prices and on-chain balances are climbing again.
Anthropic's Claude Code Adds Live, Shared Artifacts for Enterprise Teams
AI
Anthropic's Claude Code Adds Live, Shared Artifacts for Enterprise Teams
Anthropic updated Claude Code with an Artifacts feature that turns AI output into live, shared dashboards and interactive workspaces for enterprises. It pushes Claude Code from a coding assistant toward a collaboration surface where teams build and share working software artifacts in real time.
As China Looms, Taiwan Ramps Drone Production for Its Own Defense and the US Military
AI
As China Looms, Taiwan Ramps Drone Production for Its Own Defense and the US Military
Taiwan is rapidly scaling domestic drone manufacturing -- both to harden its own defenses against a potential Chinese move and to supply the US military, which is eager to diversify away from Chinese components. It's a clear marker that autonomous hardware has become a geopolitical supply-chain priority, not just a startup pitch.
New Optimization Framework Beats Claude Code and Codex by 2.5x on the Same Compute
AI
New Optimization Framework Beats Claude Code and Codex by 2.5x on the Same Compute
Researchers unveiled an AI optimization framework that reportedly outperforms leading agentic coding systems like Claude Code and Codex by 2.5x on the same compute budget. The result spotlights how much performance is still being left on the table by how agents orchestrate compute, not just which model they use.
Google Launches Its First Smart Speaker in Six Years -- Built Around Gemini
AI
Google Launches Its First Smart Speaker in Six Years -- Built Around Gemini
Google unveiled its first self-branded smart speaker in six years, built around Gemini AI, signaling a renewed push into ambient, voice-first computing as the assistant wars reignite. The hardware is a vehicle to put conversational Gemini at the center of the home.
Anthropic's Claude Code Artifacts Update Brings Live, Shared Dashboards to Enterprises
AI
Anthropic's Claude Code Artifacts Update Brings Live, Shared Dashboards to Enterprises
Anthropic updated Claude Code Artifacts to support live, shared dashboards and interactive workspaces, turning what was a single-user coding tool into a collaborative surface teams can build on together. It's a push to make Claude Code the place enterprises do work, not just generate code.
New AI Optimization Framework Claims 2.5x More Output Than Claude Code and Codex on the Same Compute
AI
New AI Optimization Framework Claims 2.5x More Output Than Claude Code and Codex on the Same Compute
Researchers introduced an optimization framework they say extracts roughly 2.5 times more useful output from coding agents like Claude Code and OpenAI's Codex while holding the compute budget fixed. If it holds up, it points to large efficiency gains sitting in orchestration rather than bigger models.
Anthropic Locks 3.5-Gigawatt Google TPU Deal as It Passes OpenAI on Revenue
AI
Anthropic Locks 3.5-Gigawatt Google TPU Deal as It Passes OpenAI on Revenue
Anthropic committed to roughly 3.5 gigawatts of Google TPU capacity through 2031 via Broadcom -- one of the largest disclosed AI compute contracts ever -- as the company's revenue run rate hit a reported $30 billion annualized, up from $14 billion in February, overtaking OpenAI. The pairing of an enormous non-Nvidia compute deal with surging revenue marks a power shift in the model race.
Anthropic Hits ~$30B Revenue Run Rate, Overtaking OpenAI
AI
Anthropic Hits ~$30B Revenue Run Rate, Overtaking OpenAI
Anthropic's annualized revenue run rate has reportedly reached about $30 billion, up from roughly $14 billion in February, putting it ahead of OpenAI and reordering the perceived leadership of the AI race. Enterprise API usage and deep partnerships with Amazon and Google are driving the surge.
'xAI Is a Failure': A Leading AI Researcher Calls Out Musk's Lab
AI3 leaders
'xAI Is a Failure': A Leading AI Researcher Calls Out Musk's Lab
One of the field's most influential researchers publicly argued that xAI has failed to become the frontier challenger Elon Musk promised three years ago, as OpenAI, Anthropic, and Google pull further ahead on both models and revenue. The critique lands as Musk's attention -- and SpaceX's newly public balance sheet -- sits elsewhere.
Google's Antigravity Replaces Gemini CLI as the Agentic Dev Default
AI
Google's Antigravity Replaces Gemini CLI as the Agentic Dev Default
Google's Antigravity -- its agentic development environment tied to the Gemini 3.5 line -- officially supersedes the Gemini CLI as of June 18, the migration deadline for free and AI Pro/Ultra users. It's Google's bid to own the agentic coding workflow as Gemini 3.5 Pro rolls out.
Google's Gemini 3.5 Pro Nears June Launch -- 2M-Token Context and Deep Think Reasoning
AI
Google's Gemini 3.5 Pro Nears June Launch -- 2M-Token Context and Deep Think Reasoning
Google's Gemini 3.5 Pro, unveiled at I/O in May, is in limited preview for select Vertex AI customers and slated for general availability in June, targeting a 2-million-token context window and Deep Think reasoning. It is Google's bid to retake the frontier as OpenAI and Anthropic battle for enterprise share.
Anthropic's Revenue Run Rate Tops $30B as It Locks Up 3.5GW of Google TPUs
AI$30B run rate
Anthropic's Revenue Run Rate Tops $30B as It Locks Up 3.5GW of Google TPUs
Anthropic said its annualized revenue run rate has passed $30B -- up from roughly $9B at the end of 2025 -- as it expands its Google/Broadcom partnership to secure 3.5 gigawatts of next-generation TPU capacity from 2027, on top of 1GW already coming online in 2026. It's one of the steepest revenue ramps and largest compute commitments in AI history.
OpenAI Releases LifeSciBench -- Frontier Models Pass Just 1 in 3 Real Science Tasks
AI
OpenAI Releases LifeSciBench -- Frontier Models Pass Just 1 in 3 Real Science Tasks
OpenAI introduced LifeSciBench, an expert-authored benchmark of 750 tasks across seven biological domains, built with 173 scientists. Even the strongest models pass only about one task in three -- a sober counterweight to claims that AI is close to autonomous scientific research.
GLM-5.2 Becomes the Leading Open-Weights Model on Artificial Analysis
AI
GLM-5.2 Becomes the Leading Open-Weights Model on Artificial Analysis
GLM-5.2 took the top spot among open-weights models on the Artificial Analysis Intelligence Index, narrowing the gap to the closed frontier labs. For builders, a state-of-the-art model you can self-host is the strongest hedge yet against the regulatory and vendor risk that just took down Fable 5.
Grok 5 Slips to Q3 -- xAI's 6-Trillion-Parameter Flagship Still in Training
AI
Grok 5 Slips to Q3 -- xAI's 6-Trillion-Parameter Flagship Still in Training
xAI's Grok 5, the rumored ~6-trillion-parameter flagship, has missed both its Q1 and Q2 2026 windows and remains in training, with full API access now looking like a Q3 story. Prediction markets price a public release before June 30 at roughly 7%, even as xAI ships smaller updates like Grok Voice and Grok Imagine Video.
Nvidia Ships Nemotron 3 Ultra -- a 550B Open Model to Seed Its Own Ecosystem
AI
Nvidia Ships Nemotron 3 Ultra -- a 550B Open Model to Seed Its Own Ecosystem
Nvidia released Nemotron 3 Ultra, a roughly 550-billion-parameter open model, extending its strategy of giving away strong models to drive demand for the hardware that runs them. As the open-weights frontier heats up, Nvidia is using openness as a moat around its silicon rather than its software.
US Government Orders Anthropic to Suspend Claude Fable 5 and Mythos 5
AI$965B valuation at risk
US Government Orders Anthropic to Suspend Claude Fable 5 and Mythos 5
Three days after launch, the US government invoked national security authorities to force-disable a live frontier model -- a regulatory first that no founder or investor had priced in. Anthropic's ~$965B valuation now carries sovereign risk.
Inside Fable 5 & Mythos 5: The Benchmarks That Made Anthropic the Model to Beat
AI
Inside Fable 5 & Mythos 5: The Benchmarks That Made Anthropic the Model to Beat
Before the shutdown, the numbers were staggering: Fable 5 topped every public reasoning benchmark and Mythos 5 became the first model to clear 90% on the agentic-coding eval. This is the capability leap the industry was reacting to -- the regulatory drama is a separate story. For builders, the question is what these models unlock, and whether the next open-weight release closes the gap.
White House EO Requires AI Developers to Share Models Before Release
AI
White House EO Requires AI Developers to Share Models Before Release
The 30-day pre-release review window gives the federal government a de facto veto on frontier model launches, as Anthropic just learned the hard way. Colorado's AI Act hitting June 30 adds state-level compliance on top.
June 2026 Becomes the Biggest AI Model Launch Month Ever
AI
June 2026 Becomes the Biggest AI Model Launch Month Ever
Gemini 3.5 Pro, Claude Fable 5/Mythos 5, rumored Sonnet 4.8, and Grok 5 -- four frontier launches in a single month signals the arms race has entered a new phase where competitive moats are measured in weeks, not quarters.
The Apple-Google AI Deal: $3B+ a Year and a New Playbook for Renting Intelligence
AI
The Apple-Google AI Deal: $3B+ a Year and a New Playbook for Renting Intelligence
Strip away the WWDC theater and this is a landmark commercial deal: Apple pays Google an estimated $3B+ annually to run Gemini behind Siri. It's the inverse of the search-default arrangement, and it sets the template for how non-lab giants will buy AI -- rent the frontier, own the distribution. Every enterprise procurement team should study the structure.