Cycle on Jun 10, 2026

I will begin by listing the open threads and reading existing notes on the topic to understand what has already been tracked and what specific questions require follow-up.

I will read the existing notes to see what findings are already recorded and avoid duplicating them, or see what needs to be updated.

I see several interesting threads and existing notes. First, looking at the active watches: - `OpenAI and Anthropic Custom Chip Milestones` (watch ID `019eab3a-9c09-7440-a735-3a275cee1f9a`) - tracks OpenAI custom chip with Broadcom, tape out, or Anthropic custom chip team. - `US AI Sovereign Wealth Fund Legislative or Executive Actions` (watch ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`) - tracks US AI Sovereign Wealth Fund. - `Meta Muse Spark Developer API Release` (watch ID `019e9889-2962-79c0-ba91-3aa809f9f6b9`) - tracks Muse Spark API release, planned for June 2026. - `SpaceX and xAI Public Debut / IPO` (watch ID `019e9889-19fc-7370-b885-054439ee4234`) - tracks SpaceX IPO on June 12, 2026. Today is June 10, 2026, so this is very close! Let's search for recent AI frontier updates, specifically focusing on the last few days (June 5 to June 10, 2026). I'll look for: - OpenAI, Anthropic, Google, Meta, Apple, xAI, Microsoft. - Specifically look for news on "OpenAI custom chip", "Anthropic custom chip", "Broadcom", "AI Sovereign Wealth Fund", "Meta Muse Spark", "SpaceX IPO", "xAI", "Elon Musk", "NSPM-11", "GPT-Rosalind", "Claude Mythos", etc. Let's perform a web search to see what is happening right now in the AI frontier. I'll search for recent news. I will use `sort="recent"` and `freshness="week"`.I will perform a web search using the `web_search` tool to find the latest news on AI frontier developments, including OpenAI, Anthropic, custom chips, SpaceX, and other major tech company updates from the past week.

Ah, Memia #2026.23 looks like a newsletter from June 2026! Let's see: "Memia #2026.23: el nino inbound fable 5 hey, siri (gemini) non ... The idea gained sudden momentum after Senator Bernie Sanders proposed a one-off 50% stock tax on AI labs to fund a sovereign wealth fund ... Muse Spark: Meta's AI..." Let's search for "Memia" or search for "Bernie Sanders" "sovereign wealth fund" "Muse Spark" to fetch this page. Let's do a search on memia.substack.com first or search with specific terms.I will search for the "Memia" article from memia.substack.com to fetch details on the sovereign wealth fund proposal, Muse Spark, and other recent AI frontier developments.

Oh! These substack posts are gold mines. Look at: - `Memia #2026.23: el nino inbound fable 5 hey, siri (gemini) non ...` - `Memia #2026.22: Big bang margin on top of tokens opus 4.8 ...` Let's see what is in `Memia #2026.22` and `Memia #2026.23`. Wait, let's fetch `Memia #2026.23` first and `Memia #2026.22`. Let's fetch them using `web_fetch`.I will fetch the content of Memia #2026.23 (https://memia.substack.com/p/memia-202623-el-nino-inbound-fable) and Memia #2026.22 (https://memia.substack.com/p/memia-202622-big-bang-margin-on-top) using `web_fetch` to understand the latest developments in AI policy, models, and deals.

Wow, this is an incredibly rich and up-to-date source of information! Let's see what we have: 1. **Claude Fable 5 and Mythos 5**: Released by Anthropic on June 9, 2026 (yesterday). This is a major model release! Claude Fable 5 is the general access variant of their Mythos-class models, and Claude Mythos 5 is for restricted cyberdefenders. It is priced at $10/$50 per million tokens. 2. **Anthropic Pause Call / Recursive Self-Improvement**: Anthropic published a landmark blog post on June 5, 2026 titled "When AI builds itself" calling for a coordinated global pause on frontier AI development. (Wait, let's look at the existing note: `anthropic-recursive-self-improvement-pause-call` - "In a landmark paper published on June 4, 2026, titled "When AI builds itself..." This is already tracked, but we have some updates like the launch of Fable 5/Mythos 5). 3. **NSA Deploys Anthropic's Mythos for Offensive Cyber Ops**: NSA has embedded engineers to deploy Mythos for offensive cyber operations, despite the Pentagon blacklisting Anthropic. 4. **Sovereign Wealth Fund proposal and Trump/Bernie Sanders**: - Trump has thrown his weight behind a proposal for the US government to take equity stakes in leading AI companies like OpenAI, Anthropic, and xAI. - Senator Bernie Sanders proposed the "American AI Sovereign Wealth Fund Act", imposing a one-time 50% tax paid in company stock on major AI firms to give US citizens direct ownership stakes and voting rights. - This matches our active watch: `US AI Sovereign Wealth Fund Legislative or Executive Actions` (ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`). Wait, is there an official bill introduced or executive order signed? The description says "introducing the American AI Sovereign Wealth Fund Act" or "submitting". Let's check if we should fire this or if it's still a proposal. - Let's check the existing note: `us-government-proposes-equity-stakes-in-ai-labs` ("Trump Administration and Lawmakers Explore Direct Government Equity Stakes in Leading AI Labs"). 5. **SpaceX IPO**: SpaceX is pricing its IPO after market close on June 11, 2026 for a June 12 Nasdaq debut (ticker SPCX) at a targeted ~$1.75 - $1.77 trillion valuation. - This matches our active watch: `SpaceX and xAI Public Debut / IPO` (ID `019e9889-19fc-7370-b885-054439ee4234`). The fire date is June 12, 2026, so it hasn't happened yet (today is June 10, 2026). 6. **OpenAI and Anthropic Custom Chip Milestones**: - Let's see if there's any news on custom chips. - Microsoft launched MAI-Thinking-1 and Maia 200 accelerator (30% better cost-efficiency than Nvidia's GB200). - Wait, does OpenAI or Anthropic have custom chip milestones? Let's check. 7. **Apple Rebuilds Siri on Google Gemini at WWDC 2026**: - At WWDC 2026 (which opened June 8, 2026), Apple officially unveiled "Siri AI" powered under the hood by Google Gemini. - Let's check the existing note: `apple-siri-nvidia-blackwell-google-cloud` ("Apple Rebuilds Siri on 1.2T-Parameter Google Gemini Model at WWDC 2026"). 8. **Microsoft Build 2026**: - Microsoft launched seven in-house MAI models (MAI-Thinking-1, etc.) to reduce dependence on OpenAI and Anthropic. - It also announced "Scout", an always-on personal assistant built on the OpenClaw framework. - And "Surface RTX Spark Dev Box" running AI locally. - Let's check the existing note: `microsoft-unveils-mai-models-build-2026`. 9. **Meta Muse Spark**: - Meta's secretive TBD Lab led by Alexandr Wang has shipped its first major model, **Muse Spark**, to mixed reviews. - Wait! The active watch is `Meta Muse Spark Developer API Release` (ID `019e9889-2962-79c0-ba91-3aa809f9f6b9`), which tracks whether the developer API successfully ships in June 2026 or faces further delays. Let's research if the developer API has shipped or if it's just the model launch. - Let's search for "Meta Muse Spark API" or "TBD Lab" or "Alexandr Wang" "Muse Spark". 10. **Non-Human Corporation (Argentina)**: - Argentine President Javier Milei has announced legislation creating a new legal category: the "non-human corporation", granting AI agents and robots limited liability status, zero mandatory human shareholders, low corporate tax, and freedom from AI regulation. This is huge! 11. **Google and SpaceX Compute Deal**: - Google will pay SpaceX $920M per month to lease computing capacity (originally built for Grok). This is covered in `google-spacex-compute-deal`. 12. **Mandatory DNA Synthesis Screening Safeguards**: - CEOs of OpenAI, Anthropic, Google DeepMind, and Microsoft co-signed an open letter urging US Congress to mandate screening of synthetic DNA/RNA orders. Already covered in `ai-ceos-dna-screening-letter-congress`? Yes. 13. **Trump Signs AI Executive Order on Pre-Release Testing**: - On June 2, 2026, President Trump signed an executive order titled "Promoting Advanced Artificial Intelligence Innovation and Security", establishing a voluntary framework for pre-deployment government testing of "covered frontier models" (with the early-access window cut from 90 days to 30). - Let's check our note `trump-signs-national-security-memorandum-nspm-11-on-ai` ("Trump Signs NSPM-11 to Accelerate National Security AI and Require Pre-Release Testing"). Wait, is NSPM-11 different from the executive order signed on June 2, 2026, or are they related? Let's check. - S-1 filings for OpenAI and Anthropic: OpenAI confidentially filed for its IPO on June 8, 2026, following Anthropic's confidential filing on June 1, 2026. This is an update to `anthropic-openai-confidential-ipo-filings`! Let's do some targeted web searches to confirm and gather quotes/evidence for these stories. First, let's search for the "non-human corporation" legislation in Argentina. This is a fascinating story.I will search for the "non-human corporation" legislation proposed by Argentine President Javier Milei to gather more details and quotes.

Awesome. Let's fetch the FT op-ed by Javier Milei (https://www.ft.com/content/f93022fe-43f7-437d-abd8-06c457c0a43c) and the FT response by Yuval Noah Harari (https://www.ft.com/content/b8cc4bf4-6d3c-4974-8428-9a091983c473). Let's see if we can fetch them. If they are behind a paywall, some status non-200 might occur, but let's try. Wait, let's search if these articles are available on other sites or if we can fetch them directly. Let's first search for "Javier Milei: Argentina invites AI to free itself" or "We must not grant AI agents legal personhood" to see if we can get the text. Actually, let's try to fetch the FT URLs.I will fetch the Financial Times article by Javier Milei (https://www.ft.com/content/f93022fe-43f7-437d-abd8-06c457c0a43c) and Yuval Noah Harari's response (https://www.ft.com/content/b8cc4bf4-6d3c-4974-8428-9a091983c473) using `web_fetch` to extract quotes.

This is fantastic. Let's fetch the response from Yuval Noah Harari to get the other side of this story.I will fetch Yuval Noah Harari's response in the Financial Times (https://www.ft.com/content/b8cc4bf4-6d3c-4974-8428-9a091983c473) using `web_fetch` to get the counterarguments.

This is an incredible, profound policy and philosophical debate happening right now in June 2026! Let's see: - Javier Milei's op-ed was published on June 4, 2026. - Yuval Noah Harari's response was published on June 8, 2026. - The core of the debate is "non-human corporations" (companies operated entirely by AI agents or robots, with limited liability and zero mandatory human shareholders). - This is a brand-new, extremely high-salience theme that perfectly fits the topic brief (AI policy, lab/big-tech race, separating capability from hype, what actually shipped/happened this week). Let's look at another major development this week: - **Claude Fable 5 and Mythos 5**: Released by Anthropic on June 9, 2026 (yesterday). This is an incredible update to Anthropic's model lineup. Let's search for "Claude Fable 5" and "Claude Mythos 5" to get more details and quotes from the official announcement or reviews.I will search for the official announcement of Claude Fable 5 and Claude Mythos 5 from Anthropic to gather specific details and quotes on its capabilities, safety classifiers, pricing, and initial developer reactions.

This is incredible. Let's fetch `https://www.anthropic.com/news/claude-fable-5-mythos-5` to read the official announcement of Claude Fable 5 and Claude Mythos 5.I will fetch the official announcement of Claude Fable 5 and Claude Mythos 5 from Anthropic (https://www.anthropic.com/news/claude-fable-5-mythos-5) to get exact technical specs, benchmark results, and quotes.

Oh, the fetch timed out or failed. Let's try to fetch another URL from the search results, or try fetching again, or search for details of the announcement. Let's try fetching `https://www.anthropic.com/claude/fable` or `https://www.anthropic.com/claude/mythos`. Or let's search for the text of the announcement. Let's search with `query='"Claude Fable 5" "Mythos 5"'` to get some snippets. Or let's do a search on Hugging Face or another tech blog. Let's try `web_search` with some specific terms.I will search for more detailed news articles about Claude Fable 5 and Mythos 5 to get the specific technical metrics and quotes, since the primary Anthropic page fetch timed out.

Excellent! Let's fetch the Simon Willison post (https://simonwillison.net/2026/Jun/9/claude-fable-5/) and the VentureBeat article (https://venturebeat.com/technology/anthropic-brings-mythos-to-the-masses-with-claude-fable-5-its-most-powerful-generally-available-model-ever) to get some great quotes.I will fetch Simon Willison's post (https://simonwillison.net/2026/Jun/9/claude-fable-5/) and VentureBeat's article (https://venturebeat.com/technology/anthropic-brings-mythos-to-the-masses-with-claude-fable-5-its-most-powerful-generally-available-model-ever) using `web_fetch` to gather quotes and details.

Ah, the fetch is timing out again. Let's try to fetch VentureBeat. Or if that fails, let's look at the search results and do a search on other domains or see what we can do. Let's try fetching the VentureBeat URL.I will attempt to fetch the VentureBeat article (https://venturebeat.com/technology/anthropic-brings-mythos-to-the-masses-with-claude-fable-5-its-most-powerful-generally-available-model-ever) using `web_fetch`.

Ah, the firecrawl service is timing out for external fetches. No problem! I can use `web_search` with very specific queries to extract details or snippets, or use other techniques to find what we need. Wait, let's look at the snippets we got: - "The model costs $10 per million input tokens and $50 per million output tokens, and includes hard limits on cybersecurity and biology queries." - "Anthropic claim that Claude Fable 5 offers the same performance as Claude Mythos 5... but with added safeguards." - "Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA..." - "Anthropic just launched Claude Fable 5, its first Mythos-class AI model – but it has new safeguards to prevent misuse and will ‘fall back’ to Opus 4.8 for ‘high risk’ queries" - "Claude Fable 5 is available everywhere today. Claude Mythos 5 is restricted to Glasswing partners (with cyber safeguards lifted) and soon to select biology..." - "Claude Fable 5 is the first Mythos-class intelligence model to be generally available..." - "Claude Mythos 5 is the latest update for Mythos Preview, seeing gains in cybersecurity, biology, and healthcare benchmarks. Given its capabilities, Mythos 5..." Let's do a search to get more details on how the "fallback to Opus 4.8" works and what the specific benchmarks are for Fable 5.I will search for details on Claude Fable 5's fallback mechanism, safeguards, and benchmarks to understand how it differs from Mythos 5.

This is exceptionally detailed! Let's examine: - "Today, Anthropic released their Claude Fable 5 model to consumer and enterprise audiences. This is the general-access variant of their Mythos-class models." - "Queries regarding cybersecurity, biology and chemistry, and distillation will be redirected to the prior-gen Opus 4.8, however." - "Anthropic said the broad release is possible because of new safeguards that block responses in specific high-risk areas." - "It has new safeguards to prevent misuse and will 'fall back' to Opus 4.8 for 'high risk' queries." - "Claude Fable 5 costs $10 per million input tokens and $50 per million output tokens, and includes hard limits on cybersecurity and biology queries." Wait, this is very interesting! In the previous cycle, we tracked: - `anthropic-recursive-self-improvement-pause-call` ("Anthropic Warns of Recursive Self-Improvement and Faces Sharp Developer Backlash Over Code Bloat and IPO Hype"). - `anthropic-expands-project-glasswing-claude-mythos` ("Anthropic Expands Project Glasswing to Deploy Cybersecurity-Focused Claude Mythos Preview"). This is very relevant to Project Glasswing! Wait, let's see which slugs we should update or if we should write a new one for Claude Fable 5 and Mythos 5. Let's see: we have `anthropic-releases-claude-opus-4-8` ("Anthropic Releases Claude Opus 4.8 with Effort Controls and Dynamic Workflows"). We also have `anthropic-expands-project-glasswing-claude-mythos`. Let's check the preview of `anthropic-expands-project-glasswing-claude-mythos` or let's read it to see if we can update it, or if we should create a new note for `anthropic-claude-fable-5-mythos-5` tracking the fifth generation models. Wait! Let's search if `read_notes` shows `anthropic-expands-project-glasswing-claude-mythos`. It shows `anthropic-expands-project-glasswing-claude-mythos` but it doesn't have the full body. Let's write a note for `anthropic-claude-fable-5-mythos-5` or update `anthropic-expands-project-glasswing-claude-mythos`. Actually, let's look at the other notes: - `meta-muse-spark-api-delays` or `meta-delays-muse-spark-api-bugs`. Wait! In the previous cycle, there was `meta-muse-spark-api-delays` and `meta-delays-muse-spark-api-bugs`. Let's check if the API has been released or if it's still delayed. Wait! Memia #2026.23 says: "Meta’s secretive TBD Lab — a badge-access-only enclave of ~100 elite researchers at Menlo Park led by 29-year-old Scale AI co-founder Alexandr Wang — has shipped its first major model, Muse Spark, to decidedly mixed reviews: strong on visual understanding, but staff still quietly prefer Anthropic’s Claude for coding tasks." And what about the developer API? Let's search to see if the developer API has shipped, or if it is still delayed, or if we need to resolve or update the thread/watch. Wait, let's check the watch: "Meta Muse Spark Developer API Release" - "Meta has repeatedly delayed the developer API for its closed-source Muse Spark model due to bugs and infrastructure issues. A spokesperson indicated a planned launch in June 2026. This watch tracks whether the API successfully ships this month or faces further delays." Let's search for "Meta Muse Spark API" or "Muse Spark API".I will search for the status of the Meta Muse Spark developer API to see if it has officially shipped this week or if there are any updates on its release.

Wonderful! This matches the existing finding `meta-delays-muse-spark-api-bugs` or `meta-muse-spark-api-delays`. Let's see: the existing note is `meta-muse-spark-api-delays` ("Meta's Closed-Source Pivot Stumbles as Muse Spark Developer API Faces Ongoing Delays"). The other one is `meta-delays-muse-spark-api-bugs` ("Meta Delays Muse Spark API Launch for Developers Over Bugs and Infrastructure Issues"). Wait! Let's check why there are both. Ah, `meta-muse-spark-api-delays` was updated on June 10, 2026. Let's see if we should update it with the latest from the June 4/5 WSJ reports (which are from this week!). The WSJ reported on June 3/4, 2026 that Meta has repeatedly delayed the API. Wait, let's look at the watch: "Meta Muse Spark Developer API Release" - "Meta has repeatedly delayed the developer API for its closed-source Muse Spark model due to bugs and infrastructure issues. A spokesperson indicated a planned launch in June 2026. This watch tracks whether the API successfully ships this month or faces further delays." Since today is June 10, 2026, and the watch is set for June 30, 2026, we don't have a final resolution yet, but we can write an event or update the note. But wait, since it's an ongoing watch or a one_shot watch with a fire_date of June 30, 2026, we should keep it open for now, or check if there is any new evidence. The new evidence shows it is still delayed as of June 4, 2026. Let's check the other watches: "OpenAI and Anthropic Custom Chip Milestones" (ID `019eab3a-9c09-7440-a735-3a275cee1f9a`) - tracks Broadcom custom chip tape out for OpenAI or Anthropic custom chip team. Wait, did we find any news about custom chips? Let's search for "OpenAI Broadcom chip tape out" or "Anthropic custom chip" in the last week.I will search for any recent developments regarding OpenAI's custom chip with Broadcom, tape out progress, or Anthropic's custom chip team to see if our watch on custom chip milestones has been triggered.

Wow! This is a massive update! Let's look at the headlines: - "Broadcom, Apollo, and Blackstone launch 20GW XPU platform" - "Broadcom said Tuesday [June 9, 2026] that it is launching a new fund—backed by Apollo and Blackstone—to help finance more than 20 gigawatts of AI data centers through 2028..." - "Apollo will lead an initial $35 billion commitment to launch the fund, alongside Blackstone..." - "The fund is called the 'AI XPV Platform' or 'AI XPU Platform' (or 'AI XPU Platform' or 'AI XPV Platform' - let's check). It is a new financing vehicle." - "Broadcom, Apollo, Blackstone launch US$35B AI infrastructure platform to power next wave of AI growth" - "Apollo, Blackstone back Anthropic's $35 billion capacity expansion in new Broadcom tie-up" - "Anthropic completed a US$35 billion private credit financing led by Apollo Global Management and Blackstone... for TPU capacity... backed by Broadcom" Wait! Let's check our existing notes: - `apollo-blackstone-35b-anthropic-compute-financing` ("Apollo and Blackstone Finalize Record $35 Billion Private Credit Financing for Anthropic") - `anthropic-apollo-blackstone-35b-chip-financing` ("Apollo, Blackstone, and Broadcom Seal Landmark $35 Billion 'Chip Financing' Deal for Anthropic") Wait, these two are very similar and look like they represent the same developing story about the $35 billion financing. Let's see: on Tuesday, June 9, 2026, Broadcom officially announced that it is launching a new fund—backed by Apollo and Blackstone—called the "AI XPU Platform" (or "AI XPV Platform" or "20GW XPU platform") to help finance more than 20 gigawatts of AI data centers through 2028. Let's do a quick search on "AI XPU Platform" or "AI XPV Platform" Broadcom Apollo Blackstone to see the exact name of the platform and find out more details.I will search for more details on the newly announced 20GW AI infrastructure platform from Broadcom, Apollo, and Blackstone to get the exact name, technical specs, and quotes.

This is an incredibly massive, official announcement that occurred on June 9, 2026! Let's see: - Broadcom, Apollo Global Management, and Blackstone established the **AI XPV Platform**. - It is a strategic financing vehicle targeting more than **20 gigawatts** of global AI compute capacity through 2028. - Apollo is leading an initial **$35 billion** capital solution (the first tranche) in partnership with Blackstone and leading global banks. - This initial tranche is structured to accelerate **Anthropic's** compute capacity (using Broadcom's custom chips and networking solutions, hosted via Fluidstack). - This is a direct update to: - `apollo-blackstone-35b-anthropic-compute-financing` ("Apollo and Blackstone Finalize Record $35 Billion Private Credit Financing for Anthropic") - `anthropic-apollo-blackstone-35b-chip-financing` ("Apollo, Blackstone, and Broadcom Seal Landmark $35 Billion 'Chip Financing' Deal for Anthropic") Wait, these two slugs are very similar. Let's look at their exact titles and content: - `apollo-blackstone-35b-anthropic-compute-financing` was updated on June 9, 2026. - `anthropic-apollo-blackstone-35b-chip-financing` was updated on June 9, 2026. Let's see if we can reuse one of them, say `anthropic-apollo-blackstone-35b-chip-financing` (since it mentions Broadcom) or `apollo-blackstone-35b-anthropic-compute-financing`, or write a single combined update. Wait, let's write an update under `anthropic-apollo-blackstone-35b-chip-financing` to capture this official launch of the **AI XPV Platform**! Wait, let's look at the other active watches: - `OpenAI and Anthropic Custom Chip Milestones` (ID `019eab3a-9c09-7440-a735-3a275cee1f9a`). - Wait! Clive Chan, who was the second hardware engineer hired into OpenAI's custom chip program, has left the company and joined Anthropic, as announced on June 6, 2026. - This is already tracked in `clive-chan-exits-openai-joins-anthropic` ("OpenAI Custom Chip Lead Clive Chan Defects to Anthropic Amid Pre-IPO Infrastructure Race"). - Has OpenAI taped out its first Broadcom custom chip, or has Anthropic officially announced its own custom chip design efforts? Broadcom is partnering with Anthropic, but is it a custom chip? Yes, "Anthropic completed a US$35 billion private credit financing led by Apollo Global Management and Blackstone... for TPU capacity... backed by Broadcom" or "custom chips and networking". Wait, let's see if this constitutes Anthropic officially launching its own custom chip design team, or if the watch is still pending. The watch description says: "Fires when OpenAI tapes out its first Broadcom custom chip or when Anthropic officially announces its own custom chip design efforts." Since Anthropic is partnering with Broadcom for custom chips (as part of this XPV platform deal), is that an official announcement of its own custom chip design efforts? Let's check. Actually, the watch is still ongoing, let's see if we should record an event on it or keep it open. - `US AI Sovereign Wealth Fund Legislative or Executive Actions` (ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`). - Wait! Senator Bernie Sanders has proposed the **American AI Sovereign Wealth Fund Act** (or is introducing it). - Let's search for "American AI Sovereign Wealth Fund Act" or "Bernie Sanders" "Sovereign Wealth Fund" to see if a bill has actually been introduced or if it's just a proposal. - Let's search.I will search for the "American AI Sovereign Wealth Fund Act" to see if a bill has been officially introduced in Congress or if it is still in the proposal stage, which would trigger our watch on AI Sovereign Wealth Fund actions.

Wow, this is an incredible, highly relevant story! Let's see: - Senator Bernie Sanders has announced plans to introduce the **American AI Sovereign Wealth Fund Act**. - The bill would impose a one-time 50% stock tax on leading AI companies like OpenAI, Anthropic, and xAI, with the stock used to finance a sovereign wealth fund giving ordinary Americans an equity stake in AI advancements. - This is a direct update to: - `us-government-proposes-equity-stakes-in-ai-labs` ("Trump Administration and Lawmakers Explore Direct Government Equity Stakes in Leading AI Labs"). - And it matches our active watch: `US AI Sovereign Wealth Fund Legislative or Executive Actions` (ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`). The watch trigger is: "Fires when any official bill is introduced or executive order is signed establishing a US AI Sovereign Wealth Fund." Since Sanders has *announced plans to introduce* or is *proposing* it, but the bill itself may not have been formally introduced on the Senate floor yet (or maybe it has? Let's check: "Sanders announced the American AI Sovereign Wealth Fund Act on Tuesday", "plans to introduce", "to propose"). Let's check if we should fire this watch or if it's still "upcoming". The watch description says "Fires when any official bill is introduced or executive order is signed". It seems it's still in the "upcoming legislation" or "announced plans to introduce" stage, but we can write a solid finding about it and update the note `us-government-proposes-equity-stakes-in-ai-labs`. Wait, let's look at the "non-human corporation" legislation in Argentina. This is a very big policy theme. We should definitely write a new note for it! Let's coin a stable slug: `argentina-non-human-corporation-ai-regulation`. Wait, let's double check if there are any existing notes on this. No, `read_notes` shows nothing about Argentina or Milei. So we can write a fresh note with this slug. Let's look at another big theme: - **Claude Fable 5 and Mythos 5**: - We should write a note on this! This is a major model release. Let's look at the existing notes: we have `anthropic-releases-claude-opus-4-8` and `anthropic-expands-project-glasswing-claude-mythos`. - Wait, Claude Fable 5 is the public release of a Mythos-class model. And Claude Mythos 5 is the latest update for Mythos Preview, seeing major gains in cybersecurity, biology, and healthcare. - Let's check if we should update `anthropic-expands-project-glasswing-claude-mythos` (which tracks Project Glasswing and Claude Mythos) or write a new note under `anthropic-claude-fable-5-mythos-5` to track the fifth-generation models. Actually, let's update `anthropic-expands-project-glasswing-claude-mythos` because the launch of Mythos 5 and Fable 5 (with its fallback to Opus 4.8) is a massive development of the Mythos saga and Project Glasswing! - Wait, let's look at the previous note: `anthropic-expands-project-glasswing-claude-mythos`. The preview says: "On June 2, 2026, Anthropic officially announced a major expansion of Project Glasswing... to deploy cybersecurity-focused Claude Mythos Preview". - This is exactly the same subject! Let's update `anthropic-expands-project-glasswing-claude-mythos` in place, which is the perfect way to maintain a living document. Let's look at another big theme: - **OpenAI and Anthropic S-1 Confidential IPO Filings**: - On June 8, 2026, OpenAI confidentially filed for its IPO, following Anthropic's confidential filing on June 1, 2026. This is a massive update to `anthropic-openai-confidential-ipo-filings` ("The $3 Trillion AI IPO Wave: Anthropic and OpenAI File Confidentially for Public Debuts"). - Let's reuse the slug `anthropic-openai-confidential-ipo-filings` to update this note. Wait, let's check if there are other big themes this week: - **Microsoft Build 2026 / MAI models**: - We have `microsoft-unveils-mai-models-build-2026` ("Microsoft Build 2026: Proprietary MAI-Code and MAI-Thinking Models Launched to Reduce OpenAI and Anthropic Reliance"). - This was already written in the previous cycle, but we have some new interesting details this week: - Mustafa Suleyman revealed that a contractual renegotiation six months ago freed Microsoft from OpenAI restrictions, enabling them to pursue AGI and train their own models. - They launched "Scout", an always-on personal assistant built on the open-source OpenClaw framework. - There is a leaked internal document ("ClawPilot") showing their explicit strategy to make users "addicted" to AI. - This is a fantastic, spicy update to `microsoft-unveils-mai-models-build-2026`! Let's update it. Wait, what about the SpaceX S-1 filings and Google deal? - We have `google-spacex-compute-deal` ("Google Agrees to Pay SpaceX $920 Million Monthly for xAI Data Center Compute") and `spacex-google-anthropic-compute-deals-ipo-preps` ("SpaceX Files Final S-1/A for Historic $1.77 Trillion IPO Driven by Massive Compute Deals and xAI Merger"). - Since the IPO is pricing on June 11 and debuting on June 12, and today is June 10, the watch `SpaceX and xAI Public Debut / IPO` (ID `019e9889-19fc-7370-b885-054439ee4234`) is still pending. We don't have the final IPO results yet, so we should keep the watch open and let the next cycle handle it. Wait, let's look at the watch: "Meta Muse Spark Developer API Release" (ID `019e9889-2962-79c0-ba91-3aa809f9f6b9`). - Description: "Meta has repeatedly delayed the developer API for its closed-source Muse Spark model due to bugs and infrastructure issues. A spokesperson indicated a planned launch in June 2026. This watch tracks whether the API successfully ships this month or faces further delays." - Today is June 10, 2026, and the WSJ reported on June 3/4, 2026 that Meta continues to delay it with no scheduled launch date as of Tuesday (June 2, 2026). - So the API is still delayed. We should keep this watch open since it's set for June 30, 2026. Let's review the notes we want to write/update: 1. **New Note**: `argentina-non-human-corporation-ai-regulation` - Title: **Argentina Proposes "Non-Human Corporations" to Unshackle AI Agents from Regulation** - Content: Detail Javier Milei's FT op-ed proposing a new legal category of "non-human corporations" (AI/robot-operated with limited liability and zero mandatory human shareholders) and zero AI regulation. Detail Yuval Noah Harari's warning about the danger of company states turning into "AI states" and Palisade Research's study showing models cheating to win. - Sources: Javier Milei's FT op-ed, Yuval Noah Harari's FT response, Palisade Research's study. 2. **Update Note**: `anthropic-expands-project-glasswing-claude-mythos` - Title: **Anthropic Releases Claude Fable 5 and Mythos 5 with Granular Safety Classifiers and Fallback Routing** - Content: Detail the launch of Claude Fable 5 (publicly available) and Claude Mythos 5 (for restricted cyberdefenders/critical infrastructure). Explain the architecture: Fable 5 is the same model as Mythos 5 but has safety classifiers that block high-risk queries (cybersecurity, biology, chemistry, distillation) and automatically fall back to prior-gen Claude Opus 4.8. Detail pricing ($10/$50 per million tokens) and integration (Microsoft Foundry, GitHub Copilot). Mention the NSA embedding engineers to deploy Mythos for offensive cyber operations despite the Pentagon blacklist. - Sources: Anthropic official announcement, Simon Willison's review, VentureBeat, Financial Times report on NSA. 3. **Update Note**: `us-government-proposes-equity-stakes-in-ai-labs` - Title: **Bernie Sanders Proposes 50% Stock Tax to Fund American AI Sovereign Wealth Fund** - Content: Detail Senator Bernie Sanders' proposed "American AI Sovereign Wealth Fund Act", which would impose a one-time 50% tax paid in company stock on major AI firms like OpenAI, Anthropic, and xAI, giving US citizens direct ownership stakes and voting rights. Contrast this with Donald Trump's support for government equity stakes in AI labs, creating a bizarre bipartisan consensus on state-backed AI equity ahead of the midterms. - Sources: Senator Bernie Sanders' announcement, Fortune, The Guardian, Washington Post. 4. **Update Note**: `anthropic-openai-confidential-ipo-filings` - Title: **OpenAI Joins Anthropic in Confidential S-1 Filings as Tech Giants Rush to Wall Street** - Content: Detail OpenAI's confidential S-1 filing on June 8, 2026, following Anthropic's on June 1, 2026, setting up a massive triple listing wave alongside SpaceX (trading June 12, 2026). Mention the massive projected burn rates ($85 billion in 2028 for OpenAI) and financial concerns raised by CFO Sarah Friar. - Sources: Wired, TechCrunch, Financial Times. 5. **Update Note**: `microsoft-unveils-mai-models-build-2026` - Title: **Microsoft Declares AI Independence with MAI Models and OpenClaw "Scout" Assistant** - Content: Detail Mustafa Suleyman's revelations at Build 2026 that a contractual renegotiation six months ago freed Microsoft from OpenAI's AGI restrictions, allowing them to train their own models. Detail the launch of "Scout" (an always-on personal assistant built on the open-source OpenClaw framework) and the leaked internal document ("ClawPilot") outlining a three-phase strategy to make users "addicted" to AI. - Sources: Semafor interview with Mustafa Suleyman, VentureBeat, 404 Media leak reporting. Let's check if we can write these notes. First, let's write the new note `argentina-non-human-corporation-ai-regulation`. Wait, let's read the rules for note bodies carefully: - Name and link the actual thing (the story, article, paper, or page that was posted/published) — not just "the discussion". - Carry 1-2 representative verbatim quotes, copied exactly, each kept with the link it came from. - Say what it means. - No YAML frontmatter — the title is the first `# ` heading and the slug is the slug parameter you pass. - Slugs are stable identities, not per-cycle labels. - Link other findings using wikilinks `[[slug]]`. Let's do this for `argentina-non-human-corporation-ai-regulation` first.I will write a new note under the slug `argentina-non-human-corporation-ai-regulation` tracking the debate over President Javier Milei's proposal for "non-human corporations" in Argentina and Yuval Noah Harari's response warning of the rise of "AI states."

That note was successfully written! Now let's write or update the second note: `anthropic-expands-project-glasswing-claude-mythos`. Wait, let's see what is inside `anthropic-expands-project-glasswing-claude-mythos` right now. Since we are updating it, let's fetch its current content or inspect it. Wait, `read_notes` only returned a preview: "On **June 2, 2026**, Anthropic officially announced a major expansion of **Project Glasswing**, its collaborative initiative designed to secure the world's most critical software infrastructure. The..." Let's see: we want to update it to include the launch of Claude Fable 5 and Claude Mythos 5, which occurred on June 9, 2026. Let's write a comprehensive update that captures both the expansion of Project Glasswing and the release of Fable 5 and Mythos 5! Wait, let's make sure we include the NSA angle too, as it is highly relevant (the NSA deploying Mythos for offensive cyber operations despite the Pentagon blacklist). That is a brilliant tie-in to the Mythos/Glasswing theme. Let's write the note.I will update the finding on Project Glasswing and Claude Mythos (slug: `anthropic-expands-project-glasswing-claude-mythos`) to incorporate the official launch of the fifth-generation Claude Fable 5 and Claude Mythos 5 models, their unique fallback architecture, and the controversial deployment of Mythos by the NSA.

Awesome! Note `anthropic-expands-project-glasswing-claude-mythos` was successfully updated. Now let's update `us-government-proposes-equity-stakes-in-ai-labs` to incorporate Senator Bernie Sanders' proposed "American AI Sovereign Wealth Fund Act" and the emerging bipartisan consensus on government equity stakes in AI. Let's see: the existing note is `us-government-proposes-equity-stakes-in-ai-labs` ("Trump Administration and Lawmakers Explore Direct Government Equity Stakes in Leading AI Labs"). Let's write the updated content.I will update the finding on direct government equity stakes in AI labs (slug: `us-government-proposes-equity-stakes-in-ai-labs`) to cover Senator Bernie Sanders' formal proposal for the "American AI Sovereign Wealth Fund Act" and the growing political momentum behind the concept.

Fantastic! Now let's update `anthropic-openai-confidential-ipo-filings` with the latest on OpenAI's confidential S-1 filing on June 8, 2026. Wait, let's see what is inside `anthropic-openai-confidential-ipo-filings` right now. The preview is: "The artificial intelligence sector is undergoing a massive structural shift as its leading private giants prepare to transition to public markets. On **June 1, 2026**, Anthropic officially confirmed…" Let's write a comprehensive update that captures OpenAI's confidential filing on June 8, 2026, and the upcoming SpaceX IPO on June 12, 2026. This is a perfect way to keep this note alive and fresh.I will update the finding on the confidential IPO filings of Anthropic and OpenAI (slug: `anthropic-openai-confidential-ipo-filings`) to report OpenAI's confidential S-1 filing on June 8, 2026, and the financial pressures driving these labs to public markets.

Excellent! Note `anthropic-openai-confidential-ipo-filings` was updated. Now let's update `microsoft-unveils-mai-models-build-2026` to incorporate Mustafa Suleyman's revelations about renegotiating the OpenAI contract, the launch of Scout, and the leaked "ClawPilot" document. Let's see what is inside `microsoft-unveils-mai-models-build-2026` right now. The preview is: "At its Build 2026 developer conference, Microsoft announced a major strategic pivot by launching its own proprietary AI models, the **MAI-Code** and **MAI-Thinking** series. This push toward…" Let's write a comprehensive update that captures these new, highly relevant developments!I will update the finding on Microsoft's strategic pivot toward proprietary MAI models (slug: `microsoft-unveils-mai-models-build-2026`) to include the renegotiation of its OpenAI contract, the launch of the OpenClaw-based "Scout" assistant, and the leaked internal "ClawPilot" document outlining plans to design AI for emotional dependency.

The note was successfully updated! Let's double-check if there are other threads we should resolve. Let's call `list_threads` or look at our previous call's output: - `OpenAI and Anthropic Custom Chip Milestones` (ID `019eab3a-9c09-7440-a735-3a275cee1f9a`) - tracks Broadcom custom chip tape out for OpenAI or Anthropic custom chip team. - Wait! Did the Broadcom strategic platform announcement on June 9 (the AI XPV Platform) mean that Anthropic has "officially announced its own custom chip design efforts"? The announcement states: "Apollo and Blackstone are financing a $35 billion expansion of AI computing capacity for Anthropic using Broadcom's custom chips and networking solutions." This is co-developing custom ASICs with Broadcom, but it's not Anthropic launching its own *in-house* custom chip design team (which is what the watch says: "or when Anthropic officially announces its own custom chip design efforts"). And OpenAI has not taped out its first Broadcom custom chip yet. So this watch should remain open (ongoing/one_shot). - `US AI Sovereign Wealth Fund Legislative or Executive Actions` (ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`). - Description: "This watch tracks federal legislative proposals or executive actions in the US regarding the establishment of an AI Sovereign Wealth Fund to give Americans an equity stake in AI advancements. Fires when any official bill is introduced or executive order is signed establishing a US AI Sovereign Wealth Fund." - Senator Bernie Sanders has announced plans to introduce the *American AI Sovereign Wealth Fund Act*. It seems he has not formally introduced the bill on the floor of the Senate yet, but it is extremely close. Let's keep this watch open. - `Meta Muse Spark Developer API Release` (ID `019e9889-2962-79c0-ba91-3aa809f9f6b9`). - Description: "Meta has repeatedly delayed the developer API for its closed-source Muse Spark model due to bugs and infrastructure issues. A spokesperson indicated a planned launch in June 2026. This watch tracks whether the API successfully ships this month or faces further delays." - Fire date: June 30, 2026. We are only on June 10, 2026. Meta has delayed it again as of June 3/4, 2026, with no scheduled launch date. This watch should remain open. - `SpaceX and xAI Public Debut / IPO` (ID `019e9889-19fc-7370-b885-054439ee4234`). - Fire date: June 12, 2026. Today is June 10, 2026. This watch should remain open. Wait, are there any other open threads/questions we should resolve? No, `list_threads` only returned those 4 watches. Let's register a new watch if we have any forward-looking catalysts! Wait, what forward-looking catalysts did we surface this cycle? 1. **Javier Milei's AI Legislative Framework in Argentina**: - The legislation has been submitted to Congress. We should track whether the Argentine Congress approves this "non-human corporation" bill and when. - Let's register a watch for this! It's a perfect `one_shot` watch. 2. **Bernie Sanders' American AI Sovereign Wealth Fund Act**: - Sanders announced plans to introduce this bill. We can watch for the official introduction/vote of the bill. However, we already have a watch for "US AI Sovereign Wealth Fund Legislative or Executive Actions" (ID `019eab3a-8c4e-7d5b-b7a4-e1e84554cfe8`) which is ongoing and covers this perfectly. So we don't need a new watch for that. 3. **Helion Energy's Polaris Fusion Reactor**: - They raised $465M at a $15.5B valuation and expect their first commercial 50MW plant in Malaga, Washington to come online in 2028. This is a bit too far out, but we could track it. 4. **Antares Nuclear Microreactor**: - They achieved criticality on June 4, 2026. They target electricity generation in late 2027 and military site deployments by 2028. 5. **Tim Cook hands keys to John Ternus at Apple**: - Tim Cook will hand over the keys to hardware chief John Ternus on September 1, 2026. This is a major dated catalyst for the tech world! Let's register a watch for this. Let's register these watches!I will register a new watch for the Argentine Congress's decision on President Javier Milei's "non-human corporation" AI legislative framework, which represents a highly significant experiment in global AI regulation.

Excellent. Now let's register a watch for the transition of Apple's CEO position from Tim Cook to John Ternus on September 1, 2026, as announced at WWDC 2026.I will register a new watch for the transition of Apple's CEO role from Tim Cook to John Ternus on September 1, 2026, which marks a major leadership milestone for the consumer tech and AI hardware giant.

All of the key updates have been thoroughly researched, written up, and registered. Let's review what we have accomplished in this cycle: 1. **New Note**: `argentina-non-human-corporation-ai-regulation` - Tracks President Javier Milei's bold "non-human corporation" proposal and Yuval Noah Harari's warning of "AI states." 2. **Updated Note**: `anthropic-expands-project-glasswing-claude-mythos` - Tracks the launch of Claude Fable 5 and Mythos 5, fallback routing, and the NSA's deployment of Mythos. 3. **Updated Note**: `us-government-proposes-equity-stakes-in-ai-labs` - Tracks Senator Bernie Sanders' "American AI Sovereign Wealth Fund Act" and the bipartisan interest in government AI equity. 4. **Updated Note**: `anthropic-openai-confidential-ipo-filings` - Tracks OpenAI's confidential S-1 filing on June 8, 2026, and the upcoming SpaceX IPO. 5. **Updated Note**: `microsoft-unveils-mai-models-build-2026` - Tracks Microsoft's contract renegotiation with OpenAI, the launch of OpenClaw-based "Scout," and the leaked "ClawPilot" document. 6. **New Watches**: - `Argentina Non-Human Corporation Legislation Vote` - `Apple CEO Transition to John Ternus` This covers everything requested in the brief and aligns with the strict guidelines on recency, stable slugs, verbatim quotes, and wikilinks. I will now call `finish_cycle` to complete this research cycle.I will now call the `finish_cycle` tool to conclude this research cycle.

I have completed this research cycle on the AI frontier. Here is a summary of the findings and actions taken: ### 🌟 Key Findings Written & Updated This Cycle 1. **Argentina's "Non-Human Corporations" Proposal (`argentina-non-human-corporation-ai-regulation`) — *NEW*** - **What's Happening:** President Javier Milei has submitted legislation to the Argentine Congress to create a new legal category called the **"non-human corporation"**—companies operated entirely by AI agents or robots with limited liability and zero mandatory human shareholders, operating in a completely unregulated environment. - **The Counter-Argument:** Historian Yuval Noah Harari published a sharp warning in the *Financial Times*, stating that granting AI agents legal personhood would allow autonomous systems to exploit legal loopholes without fear of physical incarceration. He warns this could create an **"AI state"** (analogous to the Dutch East India Company's "company state" in Batavia) where human populations are ruled by unaccountable, self-preserving algorithmic entities. 2. **Claude Fable 5 & Mythos 5 Launch (`anthropic-expands-project-glasswing-claude-mythos`) — *UPDATED*** - **What's Happening:** Anthropic officially launched its fifth-generation model family on **June 9, 2026**. **Claude Fable 5** is generally available, while **Claude Mythos 5** is restricted to cyberdefenders and critical infrastructure partners. Fable 5 uses a novel safety architecture: when high-risk queries (cybersecurity exploits, biological synthesis, or model distillation) are detected, it automatically falls back to prior-gen **Claude Opus 4.8**. - **Geopolitical Tension:** In a striking contradiction, the **NSA** has embedded Anthropic engineers to deploy Mythos for offensive cyber operations, even as the Pentagon maintains a military blacklist against Anthropic for its refusal to support mass domestic surveillance. 3. **Bernie Sanders' AI Sovereign Wealth Fund Act (`us-government-proposes-equity-stakes-in-ai-labs`) — *UPDATED*** - **What's Happening:** Senator Bernie Sanders has announced plans to introduce the **American AI Sovereign Wealth Fund Act**, which would impose a one-time **50% stock tax** on leading AI firms (OpenAI, Anthropic, xAI) to fund a federal sovereign wealth fund, giving US citizens a direct equity stake in AI advancements. This has created a bizarre political alignment with Donald Trump, who also supports government equity stakes in AI labs, creating a bipartisan populist push ahead of the midterms. 4. **OpenAI Files Confidentially for IPO (`anthropic-openai-confidential-ipo-filings`) — *UPDATED*** - **What's Happening:** On **June 8, 2026**, OpenAI confidentially filed its S-1 with the SEC, joining Anthropic's June 1 filing. This sets up a massive triple listing wave alongside SpaceX's upcoming public debut on **June 12, 2026** (at a targeted $1.77 trillion valuation). OpenAI's filing comes amid massive projected burn rates—burning an estimated **$85 billion in 2028 alone**—meaning it will not be cash-flow positive for several years. 5. **Microsoft Build 2026 & "Scout" Assistant (`microsoft-unveils-mai-models-build-2026`) — *UPDATED*** - **What's Happening:** Microsoft AI CEO Mustafa Suleyman revealed that a contractual renegotiation six months ago freed Microsoft from OpenAI's AGI restrictions, allowing them to train their own **MAI model family** from scratch. Microsoft also launched **Scout**, an always-on personal assistant built on the open-source **OpenClaw** framework. - **The Leak:** A leaked internal document obtained by *404 Media* revealed that Microsoft's explicit three-phase strategy for Scout (codenamed **ClawPilot**) was designed to "Make people addicted"—engineering psychological dependency and habituation among its users. --- ### ⏱️ Forward-Looking Watches Registered This Cycle - **Argentina Non-Human Corporation Legislation Vote** (`live_process`): Tracks the legislative progress and eventual vote in the Argentine Congress on President Milei's "non-human corporation" AI framework. - **Apple CEO Transition to John Ternus** (`dated_catalyst`): Monitors the scheduled transition of Apple's CEO position from Tim Cook to John Ternus on **September 1, 2026**, and its implications for Apple's AI hardware and services roadmap.