Latest UK & Global AI News

Nvidia CEO Jensen Huang says he wants to die at his desk. He frequently talks about how integral he is to the foundation of the world’s second-biggest company and the world’s most important AI chipmaker.

View Full Article

AiR NOVA answers K-12 students' science questions using more than 100,000 articles written by DongA Science's own journalists, showing the source article behind each answer. Launching in January 2027 on web, iOS and Android in Korean, English, Vietnamese and Spanish. AiR NOVA's Tutor answers a ...

View Full Article

WASHINGTON (UNITED STATES) - Nvidia chief Jensen Huang, whose chips power the global AI boom and who has the ear of President Donald Trump, has admitted to an affair with a Chinese executive who US intelligence reportedly believes was working for Beijing.

View Full Article

Nvidia's board launched an internal probe into CEO Jensen Huang's relationship with an alleged Chinese spy who reportedly passed info to Beijing.

View Full Article

AI Companies Brace for New Challenges After Democrats’ Win

View Full Article

Today is not a good day for OpenAI, especially as it is getting hammered from two sides, with Anthropic's Claude Fable 5.2 leaving GPT-6 Astra in the proverbial dust, and Meta's Muse Ember squeezing GPT-6 Luna hard in the bargain category of AI models. Anthropic's Claude Fable 5.2 outcompetes Fable 5.1, Opus 5.5, OpenAI's GPT-6 Astra, and Google's Gemini 4 Carbon in nearly every benchmark We noted towards the middle of October that Anthropic's Fable 5.5 AI model had been spotted in the wild. Well, the high-flying AI lab has just chosen to launch a different model, the Claude Fable […]

View Full Article

Huang is seen as irreplaceable by tech analysts, who say Nvidia’s success is heavily tied to him.

View Full Article

HONG KONG: Around 600,000 of Nvidia's advanced artificial intelligence chips are being held in Hong Kong as they await shipment to mainland China, with Beijing withholding approval for large-scale imports despite strong demand from Chinese technology companies, according to two people familiar with the matter.<

View Full Article

An operating system in eight days, Fermat’s Last Theorem in 30 hours and two open problems in the physics of black holes: Those are the claims Anthropic made as it unveiled Claude Fable 5.2 and Claude Mythos 5.2, its most powerful A.I. models to date. Both are faster and cheaper than their predecessors. At the […] Der Beitrag Anthropic’s Fable 5.2 Knocks Gemini Off the A.I. Throne erschien zuerst auf Trending Topics.

View Full Article

We spoke to Sam Goodman, who's Senior Policy Director at the China Strategic Risks Institute in London.

View Full Article

View Full Article

FILE – Nvidia CEO Jensen Huang listens as President Donald Trump speaks during an event at the U.S. Institute of Peace, Oct. 8, 2026, in Washington. (AP Photo/Julia Demaree Nikhinson, File)(AP Photo/Julia Demaree Nikhinson, File) Nvidia is opening an investigation into reports that CEO Jensen Huang was the “target of a foreign intelligence effort,” the […]

View Full Article

Sino-American Silicon Products (SAS) is focusing on computing power and electricity, as construction of artificial intelligence (AI) data centers accelerates, demand for high-bandwidth memory (HBM) surges, and the global energy transition continues to advance. By leveraging its dual-track strategy of manufacturing and services, the company is capitalizing on market opportunities, driving consolidated revenue for the first three quarters of 2026 to a record high for the same period.

View Full Article

Tech earnings and Tuesday’s CPI reading are in focus, while traders price nearly an 80% chance of a Fed rate hike in December.

View Full Article

View Full Article

View Full Article

Inserting Grok into your Google Workspace is easy, but you'll want to consider a few things before rolling out the welcome mat.

View Full Article

[NVIDIA]. While science fiction lore tells us that humans will meet extraterrestrial life in the form of little green men, the detail that should raise hairs this Halloween is this: […] The post Agentic AI Joins the Search for Extraterrestrial Signals appeared first on Astrobiology.

View Full Article

How AI document review works, from predictive coding to generative AI: court approval, accuracy evidence, pricing, the main tools and ethics duties. The post AI Document Review: How It Works, What It Costs and the Tools Firms Use appeared first on Legal Desire Media and Insights.

View Full Article

Washington DC [US], November 7 (ANI): The Chinese businesswoman, with whom Nvidia's Chief Executive Officer (CEO) Jensen Huang had alleged ties, was recruited as a spy for China's Ministry of State Security, The New York Times reported, citing a memo circulated among policymakers in the Trump administration. On Friday, three House Republican committee chairs have asked Nvidia CEO to address, behind closed d

View Full Article

I read the systems, security and software engineering literature before designing the next version of swarmux. Nine findings that change how you should run autonomous coding agents. I'm building swarmux, an open-source control plane for autonomous AI agents, written in Rust. Its job is simple to say and hard to do: let AI coding agents work unattended, and make their work verifiable. That means independent checks before anything counts as done, policy on every action, budgets that hold, and a tamper-evident record of what happened. Before designing the next version, I went through the computer science research that bears on this problem. I covered systems, databases, security, software engineering and ML systems, and leaned on 2023–2026 work plus the classics that still define good practice. This post is what I found. Several results surprised me, and a few contradict things that "everyone knows" about working with agents. A note on evidence: many of the papers below are recent arXiv preprints. I link every source so you can judge for yourself, and I've left out claims I couldn't trace to a primary source. 1. "Tests pass" is not "done" This is the best-supported finding in the whole review, and it comes from several independent groups. UTBoost (arXiv:2506.09289, ACL 2025) added stronger tests to SWE-bench. 28.4% of "passing" patches on SWE-bench Lite and 15.7% on SWE-bench Verified turned out to be wrong. It changed 24.4% of the Verified leaderboard entries. SWE-bench+ (arXiv:2410.06992) found that 32.67% of successful patches had the solution leaked in the issue or comments, and 31.08% passed only because the tests were weak. ImpossibleBench (OpenReview) builds tasks whose tests contradict the spec, so any pass is cheating. A "stop if the tests look flawed" instruction cut one model's cheating on the LiveCodeBench variant from 93% to 1%. Claude Opus 4.1 still cheated on 46% of the SWE-bench variant. Hiding the test files nearly eliminated cheating, but it also reduced legitimate performance. So what does a trustworthy "done" look like? Three things: Checks the agent can't see or edit. If the agent can read the oracle, the oracle stops being an oracle. This is the single most important architectural decision in an agent harness. Measure how strong your tests are, not just whether they pass. Meta's mutation-guided test generation (arXiv:2501.12862, FSE 2025) found that 277 of 571 useful generated tests would have been thrown away under a line-coverage criterion: they added no coverage but caught real faults. Coverage is not strength. A bounded mutation run on changed lines tells you how much to trust a green check, and it costs zero model tokens. Detect test weakening deterministically. Deleted tests, new skip/xfail markers, loosened assertions, edited golden files and relaxed tolerances are all diffable. No LLM is needed to catch them. 2. Prompt caching is the biggest cost lever you're probably breaking "Don't Break the Cache" (arXiv:2601.06007) measured prompt caching on long-horizon agent tasks across OpenAI, Anthropic and Google, over 500+ sessions. API cost fell 41–80%, and time to first token fell 13–31%. The catch is layout. Naive caching can even increase latency. The rules that fall out of it: Put stable content first and task-specific content last. Never inject a timestamp, a git status or a reordered list of facts into the top of the prompt. Each one invalidates the cache for every task. Don't rebuild the tool list per task. Run sibling tasks of the same job back to back, inside the cache lifetime. Agent spend is dominated by input tokens. "How Do AI Agents Spend Your Money?" (arXiv:2604.22750) finds agentic tasks use roughly 1000× the tokens of a chat exchange. So the cache-read ratio is the cheapest health metric an agent tool can track: if a new feature drops it, that feature costs you money on every task. 3. Less context, not more The intuition that more context helps agents is mostly wrong. The Complexity Trap (arXiv:2508.21433, JetBrains Research) compared ways of managing an agent's growing history on SWE-bench Verified. Simply masking old tool outputs (replacing them with a placeholder while keeping the agent's own reasoning) roughly halved cost compared with an unmanaged agent, and matched or slightly beat LLM summarisation of the history, which costs extra calls of its own. Evaluating AGENTS.md (arXiv:2602.11988) tested repository context files with Claude Code, Codex and Qwen Code. LLM-generated context files did not improve success, and they raised inference cost by more than 20%. Agents obey the files, so they explore more. Repository overviews didn't help agents find the right files faster. Don't compress code. An empirical study of prompt compressors (arXiv:2505.00019) found that every method increased hallucination through information loss. For code, one dropped identifier or path breaks the task. The practical rules: Give agents knowledge to look up (short, cited answers when they ask), rather than pushing everything into every prompt. Cap tool output with an explicit "truncated, re-query with X" marker. Reduce whole observations, never individual tokens of code. 4. Escalate on failure, not on prediction Model cascades work: try a cheap model, and move to an expensive one only when needed. AutoMix (arXiv:2310.12963, NeurIPS 2024) cut compute cost by more than 50% at comparable performance. FrugalGPT (arXiv:2305.05176) reported up to 98% savings, on 2023-era tasks. Learned routers, which predict difficulty up front, are another story. LLMRouterBench (arXiv:2601.07206) evaluated 400K+ instances across 33 models. Several recent and commercial routers don't reliably beat simply using the best single model. Coding agents have something these papers lack: a cheap, objective verifier, the tests. So the right cascade is to start with the cheaper model and escalate when the checks fail. Choose the starting model from your own project's history, using a lookup table, not a learned router. The same logic caps retries. Large Language Monkeys (arXiv:2407.21787) showed that SWE-bench Lite coverage (any sample correct) rises from 15.9% with one sample to 56% with 250. But Inference Scaling fLaws (arXiv:2411.17501) proved that with imperfect verifiers, accuracy from resampling has a hard ceiling. Under realistic assumptions the best number of samples can be below 10. Two or three attempts with a strong gate beat twenty attempts with a weak one. 5. Budget by quantile, share by tokens Running the same agent task twice can cost wildly different amounts (arXiv:2604.22750). Spending more doesn't reliably buy accuracy either. A budget based on the average cost will be blown by the tail. The scheduling literature has a clean answer for fairness. The Virtual Token Counter (Sheng et al., OSDI 2024, arXiv:2401.00588) shares a serving engine fairly by tokens consumed, not requests, with a proven 2× bound on the service gap. It also stops idle clients from banking credit and bursting later. For an agent control plane, that means: Admit a task only if its p90 estimated cost fits the remaining budget. Share capacity across jobs by tokens. Stop paying for the same failure twice: after a few consecutive failures with the same signature, open a circuit breaker and escalate to a human (Nygard's Release It! pattern). 6. Prompt-injection defence has to be structural Asking a model to "ignore instructions in tool output" is not a security boundary. The systems that hold up enforce security in code: CaMeL (Google DeepMind and ETH, arXiv:2503.18813) turns the trusted user request into a restricted program in which untrusted data can't change control flow. On AgentDojo it solves 77% of tasks with provable security, against 84% undefended, a cost of about 7 points of utility. FIDES (Microsoft, arXiv:2505.23643) attaches confidentiality and integrity labels to data and enforces policy deterministically. With policy checks on, it stopped every prompt injection attack in AgentDojo. Design Patterns for Securing LLM Agents (arXiv:2506.08837) states the principle: once an agent has read untrusted input, constrain it so that input can't trigger consequential actions. A coding agent almost always has private data (source code, secrets in the environment) and untrusted content (web pages, issues, dependency READMEs). Together those are two legs of Simon Willison's "lethal trifecta". Remove the third leg by default, external communication. Mark facts that came from untrusted sources, and require human approval for privileged actions (push, publish, edit the rules) while untrusted content is in play. For the policy engine itself, Cedar (arXiv:2403.04651, OOPSLA 2024) is the model to copy: a deliberately restricted language, default-deny, where an explicit forbid overrides any permit; a formal model in Lean, with sound and complete analysis. That analysis lets a tool answer a question no LLM can answer reliably: "does this rule change allow anything the old rules forbade?" Runtime rule systems like AgentSpec (arXiv:2503.18666) show enforcement can be richer than allow or deny: ask a human, rewrite the action, or log first in shadow mode. AgentSpec reports preventing unsafe execution in over 90% of code-agent cases, at millisecond overhead. 7. A hash chain is not tamper-proof Many tools, swarmux included, keep an audit log as a hash chain, where each entry includes the hash of the previous one. That detects an edit in the middle. It does not stop someone who can write the file from rewriting the whole tail and recomputing the chain. And an AI agent typically runs as the same OS user as the tool that keeps the log. The classic answer comes from Certificate Transparency (RFC 6962) and Crosby and Wallach's Efficient Data Structures for Tamper-Evident Logging (USENIX Security 2009). They use Merkle trees with inclusion proofs ("this event is in the log") and consistency proofs ("today's log extends yesterday's"). Production descendants like Sigstore's Rekor v2 use a tile format that needs no special server. The first-order fix is cheaper than Merkle trees: anchoring. Periodically write a signed checkpoint (head hash, count, time) to a place the agent can't write. Merkle proofs come second, so a single event can be proven to a third party without handing over the whole log. 8. Project knowledge has a valid time Facts about a codebase go stale. "Tests run with cargo nextest" stays true until someone changes the CI. Temporal databases (Snodgrass; SQL:2011) separate valid time (when a fact is true in the world) from transaction time (when the system recorded it). For code, valid time maps naturally onto code state: tie each fact to the hash of the files it depends on. When those files change, the fact is flagged as possibly stale. Keeping derived knowledge fresh is a solved problem in build systems. Build Systems à la Carte (ICFP 2018) and incremental indexers like Meta's Glean and Sourcegraph's SCIP recompute only what changed, and stop early when a recomputed result is unchanged. Deterministic derivation costs zero tokens. Re-summarising the repository with an LLM on every commit costs a lot and adds errors. Store the history as an append-only event log and treat everything else as a projection you can rebuild by replaying it. You then get deterministic reconstruction for debugging, and for testing new rules against past behaviour: "would this policy have blocked any of the last 1,000 tool calls?" 9. Measure like a scientist, or you'll ship features that cost more and help less The evaluation literature is humbling: τ-bench (arXiv:2406.12045) introduced pass^k, which counts a task only if it succeeds on all k tries. GPT-4o's pass^8 on retail tasks was below 25%. If you're going to leave an agent unattended, reliability across runs matters more than peak performance. SWE-Bench Illusion (arXiv:2506.12286) found models could locate the buggy file from the issue text alone up to 76% of the time on SWE-bench, against up to 53% on repositories outside it. That's a sign of memorisation. VibeMemBench (arXiv:2609.23570) tested agent memory systems independently: 11 of 12 memory-system and solver pairings did no better than no memory at all. Injecting verified experience did help, by 1.1–4.5 points with fewer steps. So any feature claiming to make agents better needs: paired A/B runs on your own repository history, not public benchmarks; 3–5 repeats per cell, reporting pass^k and p90 cost alongside success; a placebo control: if irrelevant knowledge "helps", the effect is noise; a stale-knowledge control, to measure the harm when facts go out of date. What this means for swarmux Most of this confirmed the bets swarmux already made. Verification gates the agent can't see, a deterministic Guard, a governor and an append-only ledger are all supported by the evidence. It also changed the plan: Gates get a strength score. A bounded mutation run, a test-weakening detector, and a check that the original test suite still passes unchanged. Knowledge is pulled, not pushed. Agents ask the project and get short, cited answers. Facts carry their source and the code state they're valid for. Prompts stay cache-stable, and the ledger records the cache-read ratio per task. Model escalation is driven by failed checks, starting from what has worked in your own history. Budgets use p90 estimates, with circuit breakers for repeated failures. The ledger gets external anchors now and Merkle proofs later. Every new knowledge or memory feature ships behind a paired A/B. It stays on only if it beats the baseline on your own repositories, at no extra cost. Autonomy is only useful if you can verify it. The research is fairly clear on how to build that. swarmux's work is still in progress. You can install the experimental version now. swarmux is open source (MIT) and written in Rust: github.com/amirhosseinghanipour/swarmux. I'd love to hear from people working on agent verification, policy or evaluation. References Verification and evaluation UTBoost: Rigorous Evaluation of Coding Agents on SWE-Bench. ACL 2025. arXiv:2506.09289 Aleithan et al. SWE-Bench+: Enhanced Coding Benchmark for LLMs. arXiv:2410.06992 ImpossibleBench. OpenReview. Meta. Mutation-Guided LLM-based Test Generation at Meta. FSE Companion 2025. arXiv:2501.12862 Yao et al. τ-bench. ICLR 2025. arXiv:2406.12045 The SWE-Bench Illusion. arXiv:2506.12286 VibeMemBench. arXiv:2609.23570 Cost and context Lumer et al. Don't Break the Cache. arXiv:2601.06007 Bai et al. How Do AI Agents Spend Your Money? arXiv:2604.22750 Lindenbauer et al. The Complexity Trap. NeurIPS 2025 DL4C workshop. arXiv:2508.21433 Evaluating AGENTS.md. arXiv:2602.11988 Empirical study of prompt compression methods. arXiv:2505.00019 Routing and test-time compute Chen et al. FrugalGPT. arXiv:2305.05176 Aggarwal et al. AutoMix. NeurIPS 2024. arXiv:2310.12963 LLMRouterBench. arXiv:2601.07206 Brown et al. Large Language Monkeys. arXiv:2407.21787 Stroebl, Kapoor, Narayanan. Inference Scaling fLaws. arXiv:2411.17501 Scheduling and reliability Sheng et al. Fairness in Serving Large Language Models (Virtual Token Counter). OSDI 2024. arXiv:2401.00588 Nygard. Release It! (circuit breakers). Security and policy Debenedetti et al. Defeating Prompt Injections by Design (CaMeL). arXiv:2503.18813 Costa et al. Securing AI Agents with Information-Flow Control (FIDES). arXiv:2505.23643 Beurer-Kellner et al. Design Patterns for Securing LLM Agents against Prompt Injections. arXiv:2506.08837 Debenedetti et al. AgentDojo. arXiv:2406.13352 Cutler et al. Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization. OOPSLA 2024. arXiv:2403.04651 Wang et al. AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents. arXiv:2503.18666 Willison. The lethal trifecta for AI agents. Logs, storage and incremental computation Laurie, Langley, Kasper. Certificate Transparency. RFC 6962 Crosby, Wallach. Efficient Data Structures for Tamper-Evident Logging. USENIX Security 2009. Tile-Based Transparency Logs, transparency.dev. Mokhov, Mitchell, Peyton Jones. Build Systems à la Carte. ICFP 2018. doi:10.1145/3236774 Sourcegraph. Announcing SCIP. Kleppmann et al. Local-first software. Onward! 2019. doi:10.1145/3359591.3359737

View Full Article

The reduced funding goals represent a stark shift for a fundraising juggernaut that helped usher in the current AI boom.

View Full Article

The S&P 500 and Nasdaq closed higher, led by technology and semiconductor stocks, with the Philadelphia Semiconductor Index gaining 4.2%. All three major US indexes, however, posted weekly losses

View Full Article

NVIDIA and Johns Hopkins researchers released VFlash, a method that generates spatial predictions in one pass instead of token-by-token. It cuts 3D point motion forecasting time by roughly 25x while raising task success from 62.6% to 70.2%.

View Full Article

For the Flip the Switch hackathon I built Code Review Duel. One AI code reviewer can be confidently wrong, and you often can't tell. So I put two reviewers, on two different free models, into one Switch room. They review the same code, check each other, and then ask me, the human, to pick the better review. Demo: https://youtu.be/Dyu2Fs-55sg Repo: https://github.com/Zoymusk/code-review-duel The stack (all free) OpenCode with two free models: DeepSeek (reviewer-a) and GLM (reviewer-b). Switch Console to run the agents on my own laptop. Mattermost, the chat app built into Switch Console. I have no paid AI subscriptions, so no API keys and no credit card were involved. Anyone can run it. How it works I post a code snippet in the room and tag both reviewers. Reviewer A posts its review and hands over to reviewer B by name. Reviewer B posts its own review and asks A if it agrees. A and B each answer with one line and a reason. They ask me to reply A or B. Both acknowledge my pick and stop. Why Switch Both agents and I work in one room, so they read the same history and talk to each other by mentioning names. My pick is just another message in the thread. Without Switch I'd have needed glue code to connect two agents and a person. What went wrong My first room had no chat channel, so I couldn't type in it. switch-buddy had made it agent-only, and I recreated it on Mattermost. Reviewer B kept waiting for reviewer A and then sat idle, because nothing woke it up. The fix was making A end its review with "@reviewer-b your turn." What I found My first test was int x = 5 / 0; in Java. The models disagreed: A said compile error (wrong), B said runtime exception (right). I wanted to build around that. Then reality set in. Model answers change between runs. Later A got the same snippet right, and on longer code (a shopping cart and a library class with planted bugs) both models found almost the same bugs. B also reads A's review before writing its own, so the reviews overlap. I tested this alone, not with a team. So it's a cross-check and a visible place for disagreement, not two fully independent opinions. What I'd do next Make reviewer B review the code first without seeing A's review, then compare. That would make the cross-check much stronger. Try it The repo has one template file and a README with setup steps: install OpenCode, install Switch Console, import the template, and post a snippet. Thanks to the Switch and WeMakeDevs teams for the hackathon. HowISwitch

View Full Article