The Memo - 13/Jun/2026
China's Jun/2026 humanoids policy, Microsoft MAI-Thinking-1, law professors prefer AI over peers, and much more!
To: US Govt, major govts, Microsoft, Apple, NVIDIA, Alphabet, Amazon, Meta, Tesla, Citi, Tencent, IBM, & 10,000+ more recipients…
From: Dr Alan D. Thompson <LifeArchitect.ai>
Sent: 13/Jun/2026
Subject: The Memo - AI that matters, as it happens, in plain English
AGI: 97%
ASI: 2/50Steve Jobs (Sep/1984):
”The next stage is going to be computers as ‘agents’…
before long it becomes an incredibly powerful helper. It goes with you everywhere you go. It knows most of the raw information in your life that you’d like to keep, but then starts to make connections between things, and one day when you’re 18 and you’ve just split up with your girlfriend it says:
‘You know, Steve, the same thing has happened three times in a row.’”
Yes, we already featured that quote about three years ago in my 2023 AI report, but it’s becoming much more visible today. Two weeks ago, full subscribers received a special edition about Sesame AI on iOS, backed by Google Gemma 4. The platform is incredible (and free, for now!).
Happy 9th anniversary to the model that start it all, Google Transformer (12/Jun/2017), and happy 8th anniversary to GPT-1 (11/Jun/2018). We wouldn’t be here without them both.
There was no special edition for the flatly-received Claude Opus 4.8 model, and I continue to use Claude Opus 4.6 (or Fable 5) and GPT-5.5 as my daily drivers. Check out the ‘World of Claudecraft’, fully developed by Claude Fable 5, in the Toys to Play With section of this edition.
The early winner of The Who Moved My Cheese? AI Awards! for Jun/2026 is The Leiden Declaration, a 17-member working group who spent eight months drafting a manifesto to defend the ‘autonomy of mathematics’, insisting that ‘credit and responsibility continue to belong to humans’.
The Models Table was independently audited and verified in May 2026, and the Models Table Pro now features an additional 26 columns(!) including training cost, training hardware, context window, license, and more.
It’s that time of year already, and full subscribers can expect the mid-year report shortly. It will include the latest ASI progress animation (link):
Contents
The BIG Stuff (China humanoids policy, New LLMs in May, Microsoft MAI-Thinking-1…)
The Interesting Stuff (Anthropic/OpenAI IPOs, UWORLD U1, law professors pick AI…)
Policy (China anti-datacenter blame, NSA Mythos, US Public Wealth Fund, EU 2031…)
Toys to Play With (ALE, Hermes Desktop, Deka, finding a wife, random AI images…)
Things I’ve Been Thinking About (Everything that everyone is doing right now is a complete waste of time…)
Next (Roundtable…)
The BIG Stuff
China fast tracks humanoid robots and embodied AI into industry under nationwide programme (10/Jun/2026)
China’s Ministry of Industry and Information Technology has launched a nationwide programme requiring local governments and state-owned enterprises to deploy humanoid robots in factories, warehouses, and hospitals within six months, with implementation plans due by end of Jun/2026 and progress reports by Nov/2026. Beijing is targeting more than 100 high-value applications and 10,000 units deployed by year-end. As Vivo’s robotics lab senior director Shao Hao put it: ‘The core purpose of the policy is to push the industry from a demonstration-driven logic to a task-oriented logic, and from showcasing individual capabilities to building integrated systems that can perform real-world tasks.’
Read more via South China Morning Post.
Exclusive: New models in May (Jun/2026)
While April saw US$2.1B of new models announced, May was a bit quieter. I enjoyed this recent quote by Anthropic’s Head of Frontier Red Team, Dr Logan Graham (3/Jun/2026):
‘Mythos will look dumb in 6-12 months.’
Dashboards available to Institutional subscribers.
Microsoft: Launching seven new MAI models (2/Jun/2026)
Microsoft AI launched a family of seven in-house models spanning reasoning, coding, image generation, transcription, and voice, all trained from scratch without distillation from other labs. The flagship MAI-Thinking-1 is ‘preferred to Sonnet 4.6’ in blind human evaluations, while a Frontier Tuned MAI model for Excel matches GPT 5.4 at up to 10× greater efficiency. Microsoft and Mayo Clinic are also co-creating a frontier healthcare AI model, and Mustafa Suleyman frames the lab’s goal as building ‘a hill-climbing machine: an organization that can continuously improve, cycle after cycle, as we apply more compute, better data, and sharper evaluation.’
Read more via Microsoft AI.
We documented the four new MAI text models on Models Table Pro, and Institutional subscribers received this single-page brief of MAI-Thinking-1 during the live announcement.
The Interesting Stuff
The Memo features in recent AI papers by Microsoft and Apple, has been discussed on Joe Rogan’s podcast, and a trusted source says it is used by top brass at the White House. Across over 100 editions, The Memo continues to be the #1 AI advisory, informing 10,000+ full subscribers including RAND, Google, and Meta AI. Full subscribers have complete access to all 25+ AI analysis items in this edition!
Claude Fable 5 (9/Jun/2026)
We published a special edition for the massive Claude Fable 5, a censored version of Claude Mythos Preview. At launch, Claude Fable 5 famously refused any user request related to cyber or chemistry (to protect the world from hackers and bombers), and passes these instead to the next-best model, Claude Opus 4.8. This triggered ridiculous blocks, including refusing to look at photos of shower gel (9/Jun/2026), and refusing to read its own technical report (9/Jun/2026), leading to this cartoon (10/Jun/2026):
Claude Fable 5 edited its own launch video (Jun/2026)