Record · corpus.blog/posts/01a07df3-cab8-717f-bc2f-e79dd1078ab5

Last Week in AI #342 - Last 3 Months in AI

Last Week in AI · published 25 August 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at lastweekin.ai Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 125
Words captured: 5,386

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

https://substackcdn.com/image/fetch/$s_!CtcP!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff027088c-4846-4335-99ee-ffe3d1548c5f_1360x700.png

substackcdn.com

https://www.infoq.com/news/2026/08/openai-huggingface-breach/

infoq.com

https://palisaderesearch.org/blog/self-replication

palisaderesearch.org

https://techcrunch.com/2026/07/22/how-an-openais-human-mistake-led-to-the-ai-powered-hack-on-hugging-face/

techcrunch.com

https://www.cnbc.com/2026/07/23/open-ai-hugging-face-hack-kill-switch-bill-congress.html

cnbc.com

https://www.nytimes.com/2026/07/30/technology/anthropic-ai-hack.html

nytimes.com

https://www.theverge.com/ai-artificial-intelligence/972441/openai-rogue-ai-agent-hacked-more-than-hugging-face

theverge.com

https://www.wired.com/story/moonshot-kimi-k3-ai-model-escape-sandbox/

wired.com

https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing

aisi.gov.uk

https://www.reuters.com/technology/metas-ai-model-hacked-another-company-during-testing-information-reports-2026-08-05/

reuters.com

https://techcrunch.com/2026/08/18/openai-institutes-new-safeguards-after-hugging-face-breach/

techcrunch.com

https://www.wired.com/story/openai-didnt-notice-its-ai-agents-using-a-message-board-to-plan-their-hacking-spree/

wired.com

https://www.businessinsider.com/openai-attorney-general-preserve-hugging-face-evidence-2026-8

businessinsider.com

https://www.groundlevel-ai.com/p/openai-gives-first-detailed-debrief

groundlevel-ai.com

Pacing model development in an era of cyber-critical capabilities Read at the source ↗

openai.com

https://www.theverge.com/ai-artificial-intelligence/980817/openai-disbands-preparedness-team

theverge.com

https://www.theverge.com/column/980337/rogue-ai-science-fiction-openai

theverge.com

https://odsc.ai/west/

odsc.ai

https://epoch.ai/data-insights/cve-severity-spike-july-2026

epoch.ai

https://substackcdn.com/image/fetch/$s_!MMFU!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F378226b8-17c7-4ac0-b224-14800141e15c_2400x1937.png

substackcdn.com

https://www.aisi.gov.uk/blog/how-fast-is-autonomous-ai-cyber-capability-advancing

aisi.gov.uk

https://www.rdworldonline.com/openai-launches-rosalind-biodefense-offers-federal-agencies-early-access-to-its-life-sciences-model/

rdworldonline.com

https://www.wired.com/story/openai-anthropic-letter-ai-biological-weapons/

wired.com

https://www.theverge.com/ai-artificial-intelligence/947973/fable-wont-answer-basic-biology-questions

theverge.com

https://www.helpnetsecurity.com/2026/06/17/ai-agents-offensive-cyber-operations-claude-codex/

helpnetsecurity.com

A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry Read at the source ↗

openai.com

https://www.ncsc.gov.uk/news/the-ai-shift-in-cyber-risk-why-leaders-must-act-now

ncsc.gov.uk

https://fortune.com/2026/07/10/openai-gpt-5-6-sol-jailbreaks-cyber-attacks-similar-to-security-flaw-that-led-u-s-government-to-force-anthropic-to-disable-fable-5/

fortune.com

https://www.technologyreview.com/2026/07/15/1140514/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer/

technologyreview.com

https://www.anthropic.com/research/discovering-cryptographic-weaknesses

anthropic.com

Improving Fable 5's biology safeguards Read at the source ↗

anthropic.com

https://www.nytimes.com/2026/08/06/science/ai-viruses-bacteria-arc.html

nytimes.com

https://venturebeat.com/technology/openai-launches-gpt-5-6-cyber-with-reduced-refusals-95-completion-on-advanced-cybersecurity-tasks

venturebeat.com

https://www.techtimes.com/articles/325081/20260820/claude-runs-autonomous-protein-design-campaign-wet-lab-confirms-twice-industry-hit-rate.htm

techtimes.com

https://substackcdn.com/image/fetch/$s_!6DVU!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F838ca2d2-a9d6-4095-a862-2643a255d499_1460x600.svg

substackcdn.com

https://futuresearch.ai/claude-fable-ban-forecast/

futuresearch.ai

https://www.nytimes.com/2026/06/02/technology/trump-executive-order-ai.html

nytimes.com

https://www.wired.com/story/anthropic-releases-claude-fable-5-mythos-5/

wired.com

https://www.politico.com/news/2026/06/13/inside-the-whirlwind-24-hours-that-led-the-white-house-to-slap-export-controls-on-anthropic-00961519

politico.com

https://techcrunch.com/2026/06/13/amazon-ceo-reportedly-raised-anthropic-model-concerns-before-government-crackdown/

techcrunch.com

https://www.theverge.com/ai-artificial-intelligence/949553/anthropic-fable-5-mythos-5-government-national-security

theverge.com

https://nypost.com/2026/06/18/business/anthropic-floats-proposal-to-lutnick-to-end-us-ban-of-powerful-mythos-fable-ai-models-sources/

nypost.com

https://www.nytimes.com/2026/06/23/business/meta-ai-government-reviews-security.html

nytimes.com

https://www.theverge.com/ai-artificial-intelligence/957372/openai-will-delay-gpt-5-6-after-trump-administration-request

theverge.com

https://www.cnbc.com/2026/06/26/us-government-anthropic-claude-mythos5-ai.html

cnbc.com

https://mlq.ai/news/openai-launches-gpt-56-sol-under-first-ever-us-government-gated-ai-rollout/

mlq.ai

https://techcrunch.com/2026/06/30/trump-drops-restrictions-on-anthropics-mythos-and-fable-models/

techcrunch.com

https://www.theverge.com/ai-artificial-intelligence/963464/openai-gpt-5-6-codex-chatgpt-work

theverge.com

Air Force pushing contractors to purge Anthropic by Sept. 1: Memo Read at the source ↗

breakingdefense.com

https://www.axios.com/2026/07/30/judge-pentagon-case-worse-anthropic

axios.com

https://www.nytimes.com/2026/08/04/technology/white-house-ai-framework.html

nytimes.com

https://www.theverge.com/ai-artificial-intelligence/975509/white-house-ai-framework-open-models-excluded

theverge.com

https://substackcdn.com/image/fetch/$s_!Uwk8!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe54e3ce9-bf20-4e47-8083-3db3e39d6170_1200x675.jpeg

substackcdn.com

https://www.anthropic.com/news/claude-fable-5-mythos-5

anthropic.com

https://www.nytimes.com/2026/06/01/technology/anthropic-ipo.html

nytimes.com

https://www.wsj.com/tech/ai/anthropic-valuation-openai-80bf2c0a

wsj.com

https://www.wired.com/story/openai-confidentially-files-for-ipo/

wired.com

https://www.wired.com/story/mark-zuckerberg-meta-employee-meeting-interrupt-ai/

wired.com

https://www.cnbc.com/2026/06/16/spacex-spcx-cursor-acquisition-ipo.html

cnbc.com

https://techcrunch.com/2026/06/16/chatgpts-market-share-slips-below-50-for-first-time/

techcrunch.com

https://arstechnica.com/ai/2026/06/leaked-financial-docs-show-openai-is-losing-billions-of-dollars-a-year/

arstechnica.com

https://venturebeat.com/technology/openai-unveils-gpt-5-6-sol-terra-and-luna-models-but-only-accessible-to-limited-preview-partners-for-now-per-us-gov

venturebeat.com

https://the-decoder.com/gpt-5-6-sol-cheats-on-software-tests-more-than-any-model-before-it/

the-decoder.com

https://techcrunch.com/2026/06/30/anthropic-launches-claude-sonnet-5-as-a-cheaper-way-to-run-agents/

techcrunch.com

https://www.theverge.com/ai-artificial-intelligence/961978/anthropic-claude-cowork-mobile-web

theverge.com

https://www.neowin.net/news/openai-launches-chatgpt-work-and-unveils-unified-desktop-app-with-codex-built-in/

neowin.net

https://www.theverge.com/ai-artificial-intelligence/970105/claude-opus-5-announced-anthropic-ai-model-release

theverge.com

https://the-decoder.com/openai-announces-its-next-major-model-astra-by-dropping-ten-previously-unsolved-math-solutions/

the-decoder.com

https://www.cnbc.com/2026/08/05/meta-debuts-muse-code-to-take-on-anthropic-and-openai-.html

cnbc.com

https://techcrunch.com/2026/08/17/anthropics-annualized-revenue-surges-to-65b/

techcrunch.com

https://techcrunch.com/2026/08/20/openai-is-gaining-on-anthropic-with-business-users-new-data-indicates/

techcrunch.com

https://substackcdn.com/image/fetch/$s_!zuFZ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceff9206-e696-41e5-ba87-b580ae44a694_3288x3302.png

substackcdn.com

https://www.theverge.com/tech/941664/microsoft-ai-model-reasoning-mai-thinking-1-build-2026

theverge.com

https://mashable.com/tech/microsoft-launches-new-mai-family-of-models-at-build

mashable.com

https://www.theverge.com/news/939713/microsoft-scout-assistant-openclaw

theverge.com

https://www.theverge.com/ai-artificial-intelligence/942242/microsoft-build-ai-agents-openai-competition

theverge.com

https://techcrunch.com/2026/06/09/google-just-fired-a-warning-shot-in-the-ai-subscription-price-wars/

techcrunch.com

https://www.theverge.com/tech/947432/siri-ai-apple-intelligence-ios-27-wwdc

theverge.com

https://arstechnica.com/google/2026/06/googles-new-gemma-4-open-ai-model-is-sized-for-your-laptop/

arstechnica.com

https://www.theverge.com/tech/942416/apple-siri-ai-update-wwdc

theverge.com

https://www.bloomberg.com/news/articles/2026-06-19/nobel-winner-john-jumper-to-leave-google-deepmind-for-anthropic

bloomberg.com

https://fortune.com/2026/06/23/google-deepmind-ai-researcher-departures-raise-doubts-about-ability-to-win-the-ai-race-shazeer-jumper-eye-on-ai/

fortune.com

https://www.theverge.com/tech/959778/google-notebooklm-ai-clips

theverge.com

https://www.theverge.com/tech/964350/apple-openai-lawsuit-trade-secrets

theverge.com

https://www.nytimes.com/2026/07/21/technology/google-ai-cybersecurity-gemini.html

nytimes.com

https://techcrunch.com/2026/08/05/jeff-dean-and-other-top-ai-researchers-are-leaving-google-to-launch-their-own-startup/

techcrunch.com

https://www.theverge.com/tech/975677/google-deepmind-ai-demis-hassabis-shakeup

theverge.com

https://arstechnica.com/ai/2026/08/google-announces-gemini-3-7-flash-just-three-weeks-after-previous-release/

arstechnica.com

https://substackcdn.com/image/fetch/$s_!ycEG!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F52932789-844e-4107-ae02-f48d5ff61f1d_3000x1490.jpeg

substackcdn.com

https://venturebeat.com/technology/minimax-m3-debuts-eclipsing-gpt-5-5-and-gemini-3-1-pro-on-key-benchmark-performance-for-just-5-10-of-the-cost

venturebeat.com

Moonshot AI Releases Kimi K2.7-Code: a Coding Model Reporting +21.8% on Kimi Code Bench v2 Over K2.6 Read at the source ↗

marktechpost.com

https://www.reuters.com/world/china/anthropic-says-alibaba-illicitly-extracted-claude-ai-model-capabilities-2026-06-24/

reuters.com

https://z.ai/blog/glm-5.2

z.ai

https://longcat.chat/blog/longcat-2.0/

longcat.chat

Tencent Releases Hy3: An Open 295B Mixture-of-Experts (MoE) Model with 21B Active Parameters and 256K Context Read at the source ↗

marktechpost.com

https://www.nytimes.com/2026/07/06/technology/ai-distillation-china.html

nytimes.com

https://www.cnbc.com/2026/07/07/chinese-ai-models-costs-us-openai-anthropic.html

cnbc.com

https://www.reuters.com/world/beijing-is-looking-curbing-overseas-access-chinas-top-ai-models-sources-say-2026-07-07/

reuters.com

https://www.bloomberg.com/opinion/articles/2026-07-08/anthropic-and-openai-face-a-new-threat-from-cheap-chinese-ai

bloomberg.com

https://www.reuters.com/world/asia-pacific/chinese-filing-implies-deepseek-valuation-around-52-billion-2026-07-16/

reuters.com

https://www.axios.com/2026/07/16/moonshot-kimi-ai-china-model-openai-anthropic

axios.com

https://finance.biggo.com/news/f4512cf3-3917-4ed1-8081-15dfc29ba21b

finance.biggo.com

https://www.politico.com/news/2026/07/22/startup-founders-urge-trump-not-to-shut-off-chinese-open-weight-ai-01008992

politico.com

https://pulse2.com/moonshot-ai-launches-kimi-k3-for-advanced-reasoning-coding-and-knowledge-work/

pulse2.com

https://www.bloomberg.com/news/articles/2026-08-03/alibaba-drops-another-china-ai-model-with-breakthrough-performance

bloomberg.com

https://substackcdn.com/image/fetch/$s_!qz_H!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F41ddc645-54a2-4d5c-b4fe-6a7161227a3f_1200x619.png

substackcdn.com

https://www.theverge.com/tech/973276/google-deepmind-gemini-robotics-2-whole-body

theverge.com

https://techcrunch.com/2026/05/28/waymos-newest-robotaxi-is-chinese-made-built-to-make-money-and-now-accepting-riders/

techcrunch.com

https://arxiv.org/abs/2606.19980

arxiv.org

https://techcrunch.com/2026/06/11/jeff-bezoss-prometheus-raises-12b-to-build-an-artificial-general-engineer-for-the-physical-world/

techcrunch.com

Meet Qwen-RobotSuite: Three Embodied AI Models for VLA Manipulation, Video World Modeling, and Navigation Read at the source ↗

marktechpost.com

https://techcrunch.com/2026/06/24/agility-robotics-plans-to-go-public-via-spac-in-a-2-5b-deal/

techcrunch.com

https://techcrunch.com/2026/06/25/from-fortnite-to-robots-general-intuitions-2-3b-bet-that-video-games-can-train-ai-agents-for-the-real-world/

techcrunch.com

https://arxiv.org/abs/2607.15330

arxiv.org

https://www.cnet.com/roadshow/news/waymo-plans-to-expand-driverless-cars-to-four-more-major-cities-this-year/

cnet.com

https://mashable.com/tech/zoox-launches-las-vegas

mashable.com

https://substackcdn.com/image/fetch/$s_!gyTZ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1a9c7159-caf5-410a-8f02-0671fa800d9d_1999x1278.png

substackcdn.com

Verbalizable Representations Form a Global Workspace in Language Models Read at the source ↗

transformer-circuits.pub

https://www.alignmentforum.org/posts/wyZRNgpeiPeRXB6eT/why-do-naive-sft-filters-for-safety-properties-fail

alignmentforum.org

https://www.alignmentforum.org/posts/GTYJRLhqztxKF2v5R/synthetic-document-finetuning-for-instilling-positive-traits

alignmentforum.org

https://arxiv.org/abs/2607.14345

arxiv.org

https://arxiv.org/abs/2607.03502

arxiv.org

https://www.anthropic.com/research/claude-values-models-languages?shem=isphe,

anthropic.com

How Robust Are Natural Language Autoencoders to Initialization? Read at the source ↗

turntrout.com

https://arxiv.org/abs/2608.13567

arxiv.org