• AI Valley
  • Posts
  • Microsoft AI CEO warns about Seemingly Conscious AI

Microsoft AI CEO warns about Seemingly Conscious AI

PLUS: NVIDIA and Foxconn’s humanoid robots

Together with

Howdy. It’s Barsee again.

Happy Wednesday, AI family, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

  • Microsoft AI CEO warns about Seemingly Conscious AI

  • Microsoft Excel gets an AI upgrade

  • NVIDIA and Foxconn’s humanoid robots

  • Nvidia is working on a new AI chip for China

  • Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

OUTSKILL

ChatGPT 5 is here, and it’s already changing the job market.

Some companies are cutting roles, while AI-skilled professionals earn up to $10K/month by helping automate them.

The difference? They know the frameworks and workflows most people don’t.

Join Outskill's 16-Hour AI Sprint this weekend (usually for $895) and become the AI expert companies are desperately hiring, not firing. Register now for free

Date: Saturday and Sunday, 8:30 PM - 5:30 AM (PST). 

In just 16 hours & 5 sessions, you will:

  • Build AI Agents and custom bots that handle your repetitive work and free up 20+ hours weekly

  • Learn how AI really works by learning 10+ AI tools,  LLM models and their practical use cases.

  • Learn to build websites and ship products faster, in days instead of months

  • Create professional images and videos for your business, social media, and marketing campaigns.

  • Learn new insights used by biggest giants like google, amazon, microsoft from their practitioners.

Also, unlock new bonuses in 2 days:

  • Day 1: 3000+ Prompt Bible

  • Day 2: Roadmap to make $10K/month with AI

  • Additional bonus: Your Personal AI Toolkit Builder

*This is sponsored

THROUGH THE VALLEY

Microsoft AI CEO Mustafa Suleyman has warned about a new risk he calls “Seemingly Conscious AI” (SCAI), systems that mimic memory, personality, and even subjective experiences so convincingly that people may start believing they’re sentient. In an essay, he said this could fuel cases of “AI psychosis,” where users form delusions about AI and even advocate for its rights. Suleyman argued that focusing on model welfare is “premature and dangerous,” urging companies not to market AI as conscious but to build tools for people, not as people.

Microsoft is testing a new COPILOT function in Excel that lets users generate summaries, classify data, and create tables directly inside cells using simple prompts. The feature works with existing formulas and updates results automatically as data changes. Powered by OpenAI’s GPT-4.1-mini model, it keeps inputs private but doesn’t pull from the web or company files. Microsoft warns that results may not always be accurate and has set call limits. It’s rolling out to Microsoft 365 Beta users now, with wider availability on the web coming soon.

NVIDIA is set to enter the humanoid robotics space, with longtime partner Foxconn preparing production lines ahead of a potential November reveal. The company calls this next phase “physical AI”, building on its dominance in AI training and generative AI. Reports say Foxconn will produce humanoid robots in the U.S., alongside NVIDIA’s new Blackwell Ultra AI servers, with the robots being trained for industrial use. CEO Jensen Huang has argued that physical AI could impact society even more than generative AI, making this launch a key milestone in NVIDIA’s expansion beyond computing.

Nvidia is developing a new AI chip for China, called the B30A, which could be more powerful than the H20 it currently sells there. The chip is built on the company’s Blackwell design and uses a single-die setup, likely giving it about half the power of Nvidia’s flagship B300 but still stronger than the H20. Nvidia hopes to send samples to Chinese customers as early as next month, though U.S. approval remains uncertain. Alongside it, Nvidia is also preparing the RTX6000D, a lower-spec inference chip designed to stay just under U.S. export restrictions.

OpenAI has launched a new low-cost plan in India called ChatGPT Go, priced at ₹399 ($4.57) a month, its cheapest subscription yet. The plan includes access to GPT-5, higher message limits, more image generations, file uploads, and double the memory compared to the free tier. India is OpenAI’s second-largest user base, and the rollout there comes before global expansion. The company also offers ChatGPT Plus (₹1,999 / $20) and Pro (₹19,900 / $200). Competitors like Google and Perplexity are also making aggressive pushes into India’s 800M+ internet user market.

A new Reuters/Ipsos poll shows that most Americans are uneasy about artificial intelligence. About 71% fear AI will permanently take away jobs, while 77% worry it could be used to spread political chaos through deepfakes and disinformation. Nearly half oppose using AI in military strike decisions, and many also cite concerns about energy use, AI companions replacing human relationships, and mixed impacts on education. Despite low unemployment today, Americans seem focused on the disruption AI could bring as companies invest billions in AI infrastructure.

TRENDING TOOLS

  • MovieFlo - Built by Lucasfilm & ILM vets to turn scripts into cinematic videos with seamless, pro-grade AI tools *

  • ParseBot - A web scraping/automation tool that does everything in less than 1 second

  • Firecrawl - Scrapes sites 10× faster (~1s), follows natural language like “get the blog pages”, and pulls news, images, and auto-summaries

  • Mirror - Deeply understand yourself and every relationship

  • Macaron AI - The AI that instantly gets you and cooks up mini-apps

  • Tracking AI - Tracks and ranks the IQ test scores of leading AI models

  • Qwen Image Edit - Make precise edits to any image for just $0.03 per edit

asteric (*) signifies sponsored tool

THINK PIECES / BRAIN BOOST

  • I built an AI Agent Army in n8n that completely replaced my personal assistant

  • A growing number of older Americans are turning to AI chatbots for companionship to ease loneliness by chatting, playing trivia, and exploring virtual destinations for up to five hours a day. 

  • Sam Altman on GPT-6: ‘People want memory and GPT-6 will arrive faster than the gap between GPT-4 and GPT-5’. 

THE VALLEY GEMS

What’s trending on social today:

1/ Nano Banana is most likely Gemini 2.5 with native image editing capabilities. Should be released in a day or two. Will replace Photoshop for all but professionals

2/ ChatGPT sometimes has a good sense of humour.

3/ Very impressive engineering.

4/ I hope the next GPT doesn’t fall short like GPT-5. But with so many OpenAI employees leaving for Meta, it’s hard not to be concerned.

5/ Alibaba Qwen's new image editing model is quite good.

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.