（2026.8.21~9.10）
- AI時代の衛星コンステレーション。
In the AI Era, Constellations Emerge As a New Strategic Class of Assets
For decades, military space was the business of a few select nations. Reconnaissance and secure communications rested on a handful of exquisite,
www.satellitetoday.com
- 中国AlibabaのAI、世界一のダウンロード数。
https://www.bloomberg.com/news/articles/2026-08-15/alibaba-ai-models-hit-3-billion-downloads-passing-meta-google
www.bloomberg.com
- 熱雑音を利用する「熱力学コンピューティング」。
Thermodynamic Computing Advances with Design and Training
Thermodynamic noise hinders the efficiency of computing. But what if that noise could be used as a power source instead?
newscenter.lbl.gov
- Anthropic、Mythos5より上位性能の「Model2」。
- Google/Gemini、出力の「電子透かし」をオプションで提供。
Google will now allow users to remove visible watermark from its AI generations | TechCrunch
Turning off this setting won't affect invisible benchmarks used to identify an AI generated file.
techcrunch.com
- 「AIを含むコンピューティングは新たな石油になる」
OpenAI President Greg Brockman: Compute is becoming the new oil of the AI age
Greg Brockman, OpenAI co-founder and president, joins 'Squawk Box' to discuss cybersecurity concerns around AI, OpenAI leadership changes, economics of AI, cape...
www.cnbc.com
- 秋田県に日本最大のAIデータセンター。2030年。The Japan Times 記事。
Akita eyes AI data center that could become Japan’s largest
Two firms are cooperating with the city and prefecture to make the center a hub to attract AI-related businesses and human resources.
www.japantimes.co.jp
- OpenAI、最先端AIでの学習を一時中止。セキュリティの懸念。
OpenAI Keeps Its Largest Frontier Training Run Paused Over Safety Gaps
Can frontier-AI companies keep their safeguards ahead of model capabilities well enough to justify continuing to develop and release increasingly advanced syste...
www.ainews.com
- 「中国を選ぶか?米国を選ぶか?」
The US is about to make its allies pick a side in the AI cold war
A leaked State Department letter warns 35 partner nations they cannot belong to both American and Chinese AI initiatives, turning the Pax Silica pact into a loy...
thenextweb.com
- ハリウッドとバイトダンスが和解。
MSN
www.msn.com
- イラン人ハッカー、英国の発電所をサイバー攻撃で停止させる。
https://www.telegraph.co.uk/news/2026/08/22/iranian-hackers-shut-down-uk-power-plant/www.telegraph.co.uk - Claude Mythos 5がClaude Security経由で利用可能に。
Bringing the cybersecurity capabilities of Claude Mythos 5 to more defenders | Claude by AnthropicWe’re sharing an update on our efforts to help more teams leverage frontier capabilities for cyber defense.claude.com - ホワイトハウス、2030年までに、年間1000回の打ち上げを目指す。
https://www.satellitetoday.com/government-military/2026/08/21/white-house-targets-support-for-1000-launches-per-year-by-2030/www.satellitetoday.com - OpenAIのチップ「Jalapeño」がｎVIDIAを凌ぐ
ainews.com/p/openai-jalape-o-beats-nvidia-blackwell-on-speed-and-efficiency - AIによるサイバー攻撃と防御についての共同声明。
https://openai.com/collective-cyberdefense/
openai.com
- 研究者にAIを無料提供。
オープンAI、最先端AIモデルを研究者10万人に無償提供へ
米オープンAIは7月29日、学術研究者に最先端のAIモデルを無料提供する「ChatGPT for Academic Researchers」の募集を開始した。2026年夏に1万人から開始し、2027年までに対象を10万人へ広げる。科学、数学、工学といった分野の研究や論文作成を支援する。
www.sbbit.jp
- Fable5.1 高機能・低コスト。Mythos5.1も登場。
Claude Fable 5.1の新機能
Claude Fable 5.1およびClaude Mythos 5.1における新機能、破壊的変更、機能向上の概要。
platform.claude.com
- AIで2秒以内で従来の心電図から心臓疾患を予測。
‘Superhuman’ AI tool spots heart disease in less than 2 seconds
Technology trained on millions of routine ECGs could fast-track high-risk patients for treatment
www.theguardian.com
- AI研究者が人間の研究者を超えるとき。
Automated researchers can reliably mitigate alignment failures
We had Claude autonomously train models to improve their performance on several public benchmarks that measure 10 categories of alignment failure. For all 10, C...
www.anthropic.com
- GPT-6 Astra リリース。
https://openai.com/index/gpt-6-astra/openai.com - GPT-6 AstraはAGI時代の幕開けなのか？
"Welcome to the AGI Era”: Why GPT-6 Astra Has OpenAI Talking About AGI
AGI was supposed to be obvious, but GPT-6 Astra suggests the transition may be happening before anyone agrees on what to call it.
www.ainews.com
- Muse Sparkは複雑なタスクでも止まらない。
Introducing Muse Spark 1.3
Introducing Muse Spark 1.3, with max reasoning for challenging reasoning and agentic tasks and improved real-world usability.
research.meta.ai
- AI自身がAIを開発する時代に。
https://openai.com/index/an-alien-mind/
openai.com
- フェルマーの最終定理、AIで11日で。そして公開。
Formalizing Fermat's Last Theorem
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
- 日本の大手書籍取次「日販」はAnthropicに扱いの書籍提供で合意。
「Anthropicに日販が本を売ったらしいなど」「noteやpixivなどが情プラ法大規模事業者に追加指定」など、週刊出版ニュースまとめ＆コラム #730（2026年8月30日～9月5日） | HON[.]jp News Blog
2026年8月30日～9月5日は「Anthropicに日販が本を売ったらしいなど」「noteやpixivなどが情プラ法大規模事業者に追加指定」などが話題に。広い意味での出版に関連する最新ニュースから編集長 鷹野が気になるものをピックアップ
hon.jp
- 「AIは人類を滅亡させる可能性がある」
AI Researchers: “AI Could End Humanity”—What Can We Do About the AI Race?
The researchers warning that advanced AI could become impossible to control are confronting a harder problem: how do you slow the race when every company and co...
www.ainews.com
- 企業はAIで出来るようになる未来？
Grok Bot for Enterprise
Grok Bot is now available for enterprises. Grok and Cursor Enterprise customers have free usage for the next two weeks, and can invite their whole organization,...
x.ai
コメント