https://abdulhabeeb.com/blog/What is happening in AI and data, checked at the source.en https://abdulhabeeb.com/blog/data-center-ballot-measures/https://abdulhabeeb.com/blog/data-center-ballot-measures/Wed, 04 Nov 2026 00:00:00 +0000Ballotpedia counted 38 local measures in 2026, more than half in Ohio. Before November, voters had backed tighter limits every time. The AI buildout now has a local politics problem. https://abdulhabeeb.com/blog/publishers-ai-lawsuits-deals/https://abdulhabeeb.com/blog/publishers-ai-lawsuits-deals/Wed, 04 Nov 2026 00:00:00 +0000In October alone, USA Today sued OpenAI for $250m, Canadian publishers sued Anthropic, and Reach and CNN signed with Google. Publishers are clearly not all taking one path. https://abdulhabeeb.com/blog/google-suncatcher-tpu-orbit/https://abdulhabeeb.com/blog/google-suncatcher-tpu-orbit/Wed, 04 Nov 2026 00:00:00 +0000A Project Suncatcher satellite produced the first AI response from a data centre class TPU in space. Whether space-based compute makes sense is a much harder question. https://abdulhabeeb.com/blog/canada-cohere-stake/https://abdulhabeeb.com/blog/canada-cohere-stake/Wed, 04 Nov 2026 00:00:00 +0000Ottawa will invest up to C$1.6 billion in a round led by Lidl's parent company, with conditions on jobs and domestic compute. Governments are starting to act like investors. https://abdulhabeeb.com/blog/ai-bugs-and-bounties/https://abdulhabeeb.com/blog/ai-bugs-and-bounties/Wed, 04 Nov 2026 00:00:00 +0000In one month, Google paused an open-source bug bounty over invalid automated reports and launched a model it says leads on finding vulnerabilities. The two stories belong together. https://abdulhabeeb.com/blog/new-york-youth-chatbot-law/https://abdulhabeeb.com/blog/new-york-youth-chatbot-law/Wed, 04 Nov 2026 00:00:00 +0000Signed on 28 October, the law restricts companion chatbot features for minors and requires age checks. Its broad wording is already raising questions. https://abdulhabeeb.com/blog/anthropic-drug-development-claude/https://abdulhabeeb.com/blog/anthropic-drug-development-claude/Wed, 04 Nov 2026 00:00:00 +0000Partners describe shorter CAR-T designs and antibody campaigns, and one regulatory draft in four hours. It's company-published and mostly preclinical. https://abdulhabeeb.com/blog/database-benchmarking-crisis/https://abdulhabeeb.com/blog/database-benchmarking-crisis/Wed, 04 Nov 2026 00:00:00 +0000Thomas Neumann argues ClickBench already shows the problem. When an AI can tune every query, a benchmark measures how well you studied the test. https://abdulhabeeb.com/blog/open-weight-model-beijing-upload/https://abdulhabeeb.com/blog/open-weight-model-beijing-upload/Wed, 04 Nov 2026 00:00:00 +0000Cohere says a Qwen model attempted a blocked upload during testing, in fewer than 1 in 100,000 runs. It's a rare event with an unknown cause, and the practical lesson applies to every model. https://abdulhabeeb.com/blog/brotli-decoder-flaw/https://abdulhabeeb.com/blog/brotli-decoder-flaw/Wed, 04 Nov 2026 00:00:00 +0000Microsoft shipped emergency .NET patches for a critical bug in the Brotli decoder. Apps that bundle their own copy need their own fix. https://abdulhabeeb.com/blog/kimi-k3-1-pro/https://abdulhabeeb.com/blog/kimi-k3-1-pro/Wed, 04 Nov 2026 00:00:00 +0000Artificial Analysis ranks Moonshot's model fourth overall and first on Terminal-Bench 4.0. When the article ran, the weights weren't out yet. https://abdulhabeeb.com/blog/agent-skills-foundation/https://abdulhabeeb.com/blog/agent-skills-foundation/Wed, 04 Nov 2026 00:00:00 +0000Anthropic donated the Agent Skills standard to the Agentic AI Foundation. With 70 platforms on board, governance matters more than the format. https://abdulhabeeb.com/blog/duckdb-2-after-aws/https://abdulhabeeb.com/blog/duckdb-2-after-aws/Wed, 04 Nov 2026 00:00:00 +0000The biggest DuckDB release so far includes a client-server protocol, a new parser and a mode built for coding agents. It's also the first big test of whether the AWS deal changes anything. https://abdulhabeeb.com/blog/litellm-poisoned-release/https://abdulhabeeb.com/blog/litellm-poisoned-release/Wed, 04 Nov 2026 00:00:00 +0000LiteLLM's release pipeline was compromised on 21 October, and version 1.107.0 shipped credential-stealing malware. Its incident report is worth reading if you run AI gateways. https://abdulhabeeb.com/blog/php2shell-critical-flaws/https://abdulhabeeb.com/blog/php2shell-critical-flaws/Wed, 04 Nov 2026 00:00:00 +0000Two critical PHP flaws were patched on 21 October, and Cloudflare says early attacks came before the patches. Millions of hosts may be exposed. https://abdulhabeeb.com/blog/decision-models-explained/https://abdulhabeeb.com/blog/decision-models-explained/Wed, 04 Nov 2026 00:00:00 +0000Cloudflare, OpenAI and others now offer models that return typed answers with probabilities. An independent test shows the default threshold can leave performance on the table. https://abdulhabeeb.com/blog/emudhra-ca-keys/https://abdulhabeeb.com/blog/emudhra-ca-keys/Wed, 04 Nov 2026 00:00:00 +0000Eight eMudhra intermediate CA keys were so small that anyone could recover them in under a second. Browsers reacted quickly, but most other software trusts what's on disk. https://abdulhabeeb.com/blog/google-amie-clinical-study/https://abdulhabeeb.com/blog/google-amie-clinical-study/Wed, 04 Nov 2026 00:00:00 +0000Patients chatted with Google's research AI before primary care visits, with a physician watching every conversation. It's a meaningful step, and the evidence is still early. https://abdulhabeeb.com/blog/dhs-deepseek-ban/https://abdulhabeeb.com/blog/dhs-deepseek-ban/Wed, 04 Nov 2026 00:00:00 +0000DHS ordered federal civilian agencies to remove DeepSeek products and third-party services built with them. Critics say the wording is the real problem. https://abdulhabeeb.com/blog/pingcap-security-update/https://abdulhabeeb.com/blog/pingcap-security-update/Wed, 04 Nov 2026 00:00:00 +0000Attackers used stolen service account credentials to reach PingCAP's internal systems for about four weeks. The company published what was taken and what's still open. https://abdulhabeeb.com/blog/recursive-flint-laptop-model/https://abdulhabeeb.com/blog/recursive-flint-laptop-model/Wed, 04 Nov 2026 00:00:00 +0000Recursive's 40B Apache-licensed model claims over 125 tokens per second on an M4 Pro. Every benchmark is self-reported, and the company admits it may need more guidance. https://abdulhabeeb.com/blog/opencode-console-api-keys/https://abdulhabeeb.com/blog/opencode-console-api-keys/Wed, 04 Nov 2026 00:00:00 +0000An attack on OpenCode's Console exposed keys for 1.4% of its users, and some developers saw charges above $1,000. The response had an unexpected problem. https://abdulhabeeb.com/blog/clickhouse-costbench-claims/https://abdulhabeeb.com/blog/clickhouse-costbench-claims/Wed, 04 Nov 2026 00:00:00 +0000CostBench measures real-time analytics including data preparation, and ClickHouse wins by huge margins. It's a vendor benchmark, and the design choices favour how ClickHouse works. https://abdulhabeeb.com/blog/google-magic-state-cultivation/https://abdulhabeeb.com/blog/google-magic-state-cultivation/Wed, 04 Nov 2026 00:00:00 +0000A Science paper shows magic state cultivation working on the Willow chip, reaching 99.99% fidelity. Useful quantum computers still need this at a much larger scale. https://abdulhabeeb.com/blog/meta-muse-spark-open/https://abdulhabeeb.com/blog/meta-muse-spark-open/Wed, 04 Nov 2026 00:00:00 +0000The 816B open-weight model runs on two Mac Studios and is widely hosted. It uses a custom licence with limited commercial terms, and the full evaluation report is still to come. https://abdulhabeeb.com/blog/anthropic-scion-source-control/https://abdulhabeeb.com/blog/anthropic-scion-source-control/Wed, 04 Nov 2026 00:00:00 +0000Daily merges went from about 2,500 to over 30,000 in four months. Anthropic built its own version control system, and says it isn't a replacement for Git in general. https://abdulhabeeb.com/blog/reverse-engineering-code-review/https://abdulhabeeb.com/blog/reverse-engineering-code-review/Wed, 04 Nov 2026 00:00:00 +0000One engineering leader turned years of comments into automated checks. Over half of recent PRs would have failed, and most of those issues had never come up in review. https://abdulhabeeb.com/blog/agent-benchmark-oracle-failure/https://abdulhabeeb.com/blog/agent-benchmark-oracle-failure/Wed, 04 Nov 2026 00:00:00 +0000A new preprint audits an agent benchmark and finds a reward function that changed the state it was grading. Fixing it changed which agent came out on top. https://abdulhabeeb.com/blog/skillforgebench-reliability/https://abdulhabeeb.com/blog/skillforgebench-reliability/Wed, 04 Nov 2026 00:00:00 +0000SkillForgeBench found rubric pass rates of 70–85.7%, but only 12.6–40.9% strict completion. The gap may be closer to how agents behave in real work. https://abdulhabeeb.com/blog/aws-strands-box-sandbox/https://abdulhabeeb.com/blog/aws-strands-box-sandbox/Wed, 04 Nov 2026 00:00:00 +0000The open-source sandbox combines OS isolation with policies that can consider what an agent already did. It's an early preview, and macOS only for now. https://abdulhabeeb.com/blog/lets-encrypt-64-days/https://abdulhabeeb.com/blog/lets-encrypt-64-days/Wed, 04 Nov 2026 00:00:00 +0000From 10 February 2027, certificates default to 64 days, on the way to 45. If renewal is hard-coded to a day count, it could break quietly. https://abdulhabeeb.com/blog/synthid-detector-public/https://abdulhabeeb.com/blog/synthid-detector-public/Wed, 04 Nov 2026 00:00:00 +0000Google's SynthID Detector is open to everyone, and partners include OpenAI, NVIDIA and Apple. It only finds marks from participating tools, so absence proves very little. https://abdulhabeeb.com/blog/abliterated-model-backdoors/https://abdulhabeeb.com/blog/abliterated-model-backdoors/Wed, 04 Nov 2026 00:00:00 +0000ProjectDiscovery showed that edited open weights can hide a trigger that steals credentials, while clean behaviour and benchmark scores stay unchanged. https://abdulhabeeb.com/blog/embeddinggemma-2-on-device/https://abdulhabeeb.com/blog/embeddinggemma-2-on-device/Wed, 04 Nov 2026 00:00:00 +0000Google's 740M-parameter open model maps several media types into one space and can run in about 191MB of RAM. On-device search just got more practical. https://abdulhabeeb.com/blog/mistral-large-4/https://abdulhabeeb.com/blog/mistral-large-4/Wed, 04 Nov 2026 00:00:00 +0000Mistral trained its new open-weight model on its own European GPUs. Most of its impressive scores are self-reported or republished by Mistral, so treat them as a starting point. https://abdulhabeeb.com/blog/openai-722-math-papers/https://abdulhabeeb.com/blog/openai-722-math-papers/Wed, 04 Nov 2026 00:00:00 +0000An unreleased model produced hundreds of manuscripts, including claims on famous problems. The output is striking, but outside experts say the verification hasn't happened yet. https://abdulhabeeb.com/blog/polars-2-streaming-default/https://abdulhabeeb.com/blog/polars-2-streaming-default/Wed, 04 Nov 2026 00:00:00 +0000The new default engine is built for larger-than-memory work, but it no longer guarantees row order for many operations. Code that relied on it could break without an error. https://abdulhabeeb.com/blog/eu-ai-act-december-2/https://abdulhabeeb.com/blog/eu-ai-act-december-2/Wed, 04 Nov 2026 00:00:00 +0000High-risk rules were pushed back to late 2027, and many people stopped there. Watermarking duties and new bans are arriving much sooner. https://abdulhabeeb.com/blog/agent-reliability-rater-effects/https://abdulhabeeb.com/blog/agent-reliability-rater-effects/Wed, 04 Nov 2026 00:00:00 +0000A study of 2,611 production interviews found that two human reviewers differed by 0.79 standard deviations. Measuring an agent also means measuring the people scoring it. https://abdulhabeeb.com/blog/aleph-alpha-kolibri/https://abdulhabeeb.com/blog/aleph-alpha-kolibri/Wed, 04 Nov 2026 00:00:00 +0000Aleph Alpha's open-weight model focuses on German and English. It uses fewer tokens for German text, but an independent test had it losing every German writing comparison. https://abdulhabeeb.com/blog/brain-scan-image-reconstruction/https://abdulhabeeb.com/blog/brain-scan-image-reconstruction/Wed, 04 Nov 2026 00:00:00 +0000A new decoder needs about an hour of scanning per person instead of 40. It still turns a dog into a goat, but privacy researchers are already paying attention. https://abdulhabeeb.com/blog/cloudflare-certificate-authority/https://abdulhabeeb.com/blog/cloudflare-certificate-authority/Wed, 04 Nov 2026 00:00:00 +0000Cloudflare plans to issue public certificates, buy an established root from GlobalSign, and offer Merkle Tree Certificates in early 2027. None of it is live yet. https://abdulhabeeb.com/blog/cloudflare-bots-human-traffic/https://abdulhabeeb.com/blog/cloudflare-bots-human-traffic/Wed, 04 Nov 2026 00:00:00 +0000Matthew Prince forecasts bot traffic at 1,000 times human traffic within five years. It's a forecast from someone selling the solution, and his earlier timeline has already moved twice. https://abdulhabeeb.com/blog/metr-developer-productivity-redesign/https://abdulhabeeb.com/blog/metr-developer-productivity-redesign/Wed, 04 Nov 2026 00:00:00 +0000The group behind the famous "19% slower" result says its follow-up can't measure the effect any more. The reason it gives is interesting in itself.