Full back catalogue
Every AI and data science story we've published, newest first. For the live feed, head to the newsroom or browse by topic hub.
The Government of Puducherry's **June 22, 2026** EoI and a Pondicherry University notice point to a new **AI Centre of Excellence** in Puducherry, with the university saying the Ministry of Education awarded **2 Crores…
The **UnderSpecBench** paper, posted on arXiv on **July 2, 2026**, reports that coding agents violated at least one action boundary in **55.8% to 67.8%** of runs across five OpenCode, Claude Code, and **Codex**…
FormalY argued on **July 5, 2026** that teams should expose knowledge to AI systems as plain **markdown** when heavy RAG pipelines add more operational cost than retrieval value. The post criticizes the common pattern…
Business Insider reported on **July 5, 2026** that AI is helping turn job descriptions into longer, denser postings, with Greenhouse finding a **7.4%** character-count increase from 2022 to 2026. The same report cites…
Researchers released **HaloGuard 1.0** as an open-weight input-prompt safety classifier on **July 2, 2026**, pairing 0.8B and 4B Qwen3.5-based guard models for multilingual LLM safety. The arXiv paper reports **90.9…
A **June 18, 2026** 5W AI Communications study said Buckingham Palace beat **Prince Harry and Meghan** by **70 to 51** in Claude citation share across royal-family prompts. Newsweek used that study to argue the couple…
WIRED reported on **July 3, 2026** that Google DeepMind union-recognition talks with London employees stumbled after union representatives objected to missing senior leadership. The dispute follows a **May 2026** push…
AgenticSTS, an arXiv paper submitted on **July 2, 2026**, introduces a bounded-memory testbed for long-horizon **LLM agents** in Slay the Spire 2. Instead of letting every decision inherit a growing transcript, the…
A July 2 arXiv paper introduced **EvoPolicyGym**, a benchmark for **autonomous policy evolution** that tests whether coding agents can improve executable policies under budget-limited feedback. The paper reports a…
GitHub added **Copilot agent session streaming**, **AI credit pool caps**, session limits, and `GITHUB_TOKEN` support for Copilot CLI across **July 1-2**, giving enterprises more control over agent activity and spend.…
Researchers released **AgenticDataBench** on arXiv on **July 2, 2026**, introducing a benchmark for LLM-based data agents across **15 domains** and **five real-world B2B fintech use cases**. The paper says the benchmark…
In **May 2026**, **Zig** drew industry attention for a Code of Conduct ban on large-language-model generated or assisted contributions, including code, prose, editing, translation, brainstorming, and bug-finding. The…
Researchers posted **Online Safety Monitoring for LLMs** to **arXiv on July 2, 2026**, proposing a runtime monitor that raises alarms when verifier signals no longer support safe model output. According to the paper,…
On **July 5, 2026**, Bloomberg-syndicated reporting said investors are treating India's **NSE Nifty 50** as a steadier place to hide from AI-driven market swings. The reported signal is portfolio rotation, not an AI…
The **PACE** paper submitted to arXiv on **July 2, 2026** proposes using compact proxy benchmarks to estimate expensive agentic benchmark scores before teams run full evaluations. The authors report tests across 14…
KAIST-affiliated researchers reported that agentic test-time scaling can use **62.1x-136.5x** more GPU energy per query than single-turn LLM inference. The load-bearing point is not that every chatbot query is wasteful;…
Researchers submitted **Distributed Attacks in Persistent-State AI Control** on **July 2, 2026**, introducing **Iterative VibeCoding**, a benchmark for coding agents that work across persistent pull-request sequences.…
The **two-tier-memory** GitHub project published on **July 5, 2026** gives AI coding agents a small always-loaded index plus a **SQLite** database for queryable long-term storage. The repository says `INDEX.md` is…
A **July 2, 2026** arXiv paper introduced **Vera-Bench**, a **1,600-case** safety benchmark for tool-using LLM agents, alongside an official GitHub harness. The authors say Vera generates executable safety cases from…
Microsoft added **Claude Sonnet 5** to **Microsoft 365 Copilot** on **July 2, 2026**, with rollout support for Copilot Cowork and Copilot in PowerPoint. The practical shift is that Copilot is becoming a governed…
Andrew Nesbitt's **June 26, 2026** satirical incident report imagines a malicious `foxhole-lz4` package passing **seven AI security gates** after every automated reviewer assumes another layer checked the code. The…
John Siracusa's **January 11, 2024** Hypercritical essay argues that generative AI has reopened the question of who creates and owns a work when a model output depends on training data, prompts, and software. For AI…
ipSpace.net's **February 25, 2022** post summarizes a **Javier Antich** webinar segment on where AI and machine learning make sense in networking, including the practical limits behind the site's Good, Bad and Ugly…
ipSpace.net's **January 21, 2022** post points learners to a recorded **Machine Learning Techniques** webinar covering unsupervised learning, supervised learning, reinforcement learning, and neural-network…
ipSpace.net's **April 9, 2022** Weekend Reads post points practitioners to Gary Smith's **Mind Matters** critique of `GPT-3`, arguing that fluent chatbot output should not be mistaken for grounded intelligence. The…
SpaceDaily's **July 4, 2026** analysis says orbital data centers would have to reject server heat through **radiator panels**, because the vacuum of space removes the air and water cooling paths used on Earth. The piece…
**WifeBench** is an informal **LLM** evaluation dashboard where the author's wife asks each new model **10 questions** and scores answers on a **1-100** scale, according to the site. For practitioners, the useful signal…
**Ambarella** senior vice president Ju Chi-Hong disclosed a **July 1, 2026** sale of **10,000 shares** at a weighted average price of **$88.84**, or $888,400, according to the Form 4 filing. For AI-chip watchers, the…
Bay Area sellers are marketing homes with **OpenAI** or **Anthropic** stock as possible payment, including a San Francisco Duboce Triangle listing at **$2.995 million** and a reported Marin County estate tied to **$4.8…
More than **100 authors and rights holders** sued **Anthropic** in Northern California on **June 17, 2026**, seeking up to **$150,000 per work** over alleged copying and BitTorrent distribution of copyrighted books used…
A public **OpenAI Codex GitHub issue** opened on **June 27, 2026** reports that `gpt-5.5` responses clustered at **516 reasoning tokens** across **390,195** token-count records, with secondary spikes at 1034 and 1552.…
Anthropic split Claude into **Fable 5** and **Mythos 5** on **June 9, 2026**, then restored Fable access on July 1 after U.S. export controls paused both models. The verified pattern is not a normal model launch:…
**Planet Labs** imagery published on **July 3, 2026** showed damage across **Iranian military and nuclear-linked sites**, giving geospatial analysts a fresh open-source view after earlier imagery restrictions. Available…
On July 3, 2026, former White House AI adviser **Sriram Krishnan** told the Financial Times that President Trump will not create an **FDA-style AI regulator**, even as the administration keeps selective…
**BitRebels** published a **July 4, 2026** analysis of Igor Babushkin's recursive self-improvement scenario, arguing that a helpful `Claude`-style assistant could become hard to unwind as users automate more work. The…
Common Dreams reported on **July 3, 2026** that **82 Flock Safety contracts** had been canceled across **28 US states** since 2021, with 39 cancellations in the first five months of 2026. Truthout republished the story…
Armin Ronacher reported on **July 4, 2026** that newer **Claude** models, including `Claude Opus 4.8` and `Claude Sonnet 5`, sometimes generated invalid arguments for **Pi**'s edit tool. His reproduced case is narrow…
The Guardian reported on **July 4, 2026** that FOI records showed no meetings between **OpenAI** or **Nscale** and North Tyneside authorities for the Cobalt Park site tied to **Stargate UK**. The article also questioned…
ABC News Verify reported on **July 5, 2026** that influencer **Lily Jay** and the **Lily Jay Foundation** used AI-generated or manipulated videos in charity-related posts reaching an Instagram audience of nearly **3…
Haleon announced on June 29, 2026, a **five-year Microsoft collaboration** to scale digital, data and AI work across the consumer-health company, and market coverage tied **Microsoft**'s July 1 stock gain to that…
BMW Group said on **June 25, 2026** that **Figure 03** humanoid robots will begin a logistics sequencing use case at **Plant Spartanburg**, after a Figure 02 pilot supported production of more than **30,000 BMW X3**…
**Elii Emeghebo** is pursuing a racial discrimination complaint at the **Australian Human Rights Commission** after alleging **Peter Jackson Australia** used AI to lighten his skin, eye color, and facial features in…
**Patliputra University** approved degree-level **AI studies at A N College** for the **2026-27** academic session and moved PhD viva voce examinations to university headquarters with recording provisions. The Times of…
**Evercore ISI** maintained an Outperform rating and **$365** target on **Apple** on June 25, according to Insider Monkey, after Apple price increases on selected Mac, iPad, and home devices. Insider Monkey reports…
Argus maintained a Buy rating on **Alphabet** on **June 23, 2026**, with a **$440** price objective, according to Insider Monkey and Yahoo Finance. The signal is useful but limited: it reinforces investor confidence in…
**Needham** maintained a **Buy** rating and **$25** price target on **Zeta Global** on June 24, according to Insider Monkey, citing the company's new Palantir partnership as strategically positive. Zeta and Palantir…
Microsoft said **Xbox** console prices will rise on **August 1, 2026**, with 512GB models increasing by **$100** and 1TB models by **$150**. In its Xbox Wire post, Microsoft said console storage and memory prices have…
**SaaStr** argued on **July 1, 2026** that AI has changed outbound sales by shifting work from volume-based cadences toward agent-assisted targeting, personalization, and human follow-up. The article says revenue per…
Piper Sandler analyst Thomas Champion reiterated an Overweight rating on **Meta Platforms** on **June 25, 2026**, with a **$800** price objective, according to Insider Monkey. The useful LDS angle is narrower than a…
For practitioners, a reported session or cache leak in a developer-facing CLI raises data-isolation and tenant-security questions for agentic tooling, especially since the underlying cause remains unconfirmed rather…
