In partnership with
🌟 Editor's Note: Recapping the AI landscape from 10/28/25 - 11/03/25.
🎇✅ Welcoming Thoughts
Welcome to the 17th edition of NoahonAI.
What’s included: company moves, a weekly winner, AI industry impacts, practical use cases, and more.
Checking in on Tuesday night and there’s already been more craziness today than the entirety of last week.
Humanoid robots are here and they’re awesome.
Thermocomputing may change the world. See Startup Spotlight.
Microsoft CEO went on TBPN rather than a major network to talk OpenAI deal, times are a-changing.
Did I just see ‘Apple’ in AI news, I must’ve misread it.
I’m running into a bug with GPT voice chat where it hears it’s own response and thinks that it’s me talking… not ideal.
Apparently OpenAI and Anthropic almost merged last year when Sam Altman was briefly ousted as CEO 🤯.
NVIDIA CEO said that plumbers and electricians will be needed by the hundreds of thousands in his future working world vision.
Grok and Meta, please provide more interesting news.
Some false reports out there that OpenAI can no longer deal with anything health or legal related. Misinformation.
Heavy week of action at the top of the AI race.
Let’s get started—plenty to cover this week.
👑 This Week’s Winner: NVIDIA
Absolute battle back and forth this week and NVIDIA just beats out OpenAI. In a week where OpenAI restructured to prepare for the next 5-10 years, NVIDIA needed a dominant performance across the board to win, and thats what they got.
NVIDIA started the week by becoming the first company in history to hit a $5 TRILLION DOLLAR market cap. A number which has since increased up to $5.12T. That was just the beginning.
Revealed $500B in Orders: This likely encapsulates the next five to six quarters and includes AI chips / infrastructure. Underscores surging global demand from companies, governments, and more. Whoever can buy is buying.
Partnered with SK Group: NVIDIA is building an AI factory in South Korea powered by 50,000+ GPUs, supporting semiconductor design, robotics, digital twins, and intelligent AI agents. The first phase will finish by late 2027.
New AI Supercomputers: Built for the U.S. Department of Energy, including one system with 100,000 Blackwell GPUs powering next-gen national lab research.
Partnered with Samsung: Building another “AI factory” that brings AI into chip manufacturing with Samsung technology, boosting lithography speed, and integrating robotics.
After going through that again, it’s clear NVIDIA is this week’s winner. Incredibly strong showing, providing not just shovels for the gold rush, but hammers, drills and everything needed across the board.
From Top to Bottom: Open AI, Google Gemini, xAI, Meta AI, Anthropic, NVIDIA.
⬇️ The Rest of the Field
Who’s moving, who’s stalling, and who’s climbing: Ordered by production this week.
🟢 OpenAI // ChatGPT
OpenAI For Profit: OpenAI is transitioning its for-profit arm into a public-benefit corporation, designed to attract investors while maintaining its mission-driven oversight. $1T IPO expected for 2026. This is HUGE for OpenAI. They need money and this will bring it. Still no equity for Altman in restructure.
$38B AWS Deal: OpenAI signed a 7-year agreement with AWS for large-scale cloud compute access, reducing reliance on Microsoft Azure. Still using NVIDIA GPU’s through Amazon’s cloud computing. More compute.
New Structure New Roadmap: CEO Sam Altman outlined plans for OpenAI’s models to evolve into autonomous “research agents” by 2028, framing AGI as a gradual, multi-year progression. ChatGPT is just the start. We’re still early.
🟠 Anthropic // Claude
Cognizant Partnership: Cognizant became one of Anthropic’s largest customers, deploying Claude to 350k employees and co-selling it to clients to expand enterprise adoption. Anthropic continues to dominate the enterprise space.
Claude Meet Claude: Anthropic’s new paper claims Claude shows limited introspective awareness of its internal states, though researchers stress it’s not true consciousness. Basically the model noted when researchers added data into its ‘thinking’ process. Still far from self-awareness but fascinating nonetheless. Could be useful to solve things like prompt injection.
Tokyo Expansion: Anthropic opened a Tokyo office and signed a deal with Japan’s AI Safety Institute to collaborate on evaluation standards and safe AI development. Expected. Good expansion move.
🟣 Google // Gemini
Apple x Gemini: Apple is reportedly in talks with Google to power a new Siri using Gemini, with rollout expected in iOS 26 (2026). They previously hinted at this. Would be a much needed upgrade!
Gemini Maps: Gemini is being added to Google Maps, enabling conversational navigation and trip planning. Currently in beta. API just released so this is expected. Excited to see what it will look like.
Gemini Defense: Lockheed Martin will integrate Gemini into secure, air-gapped defense systems for data analysis and R&D, signaling Google’s deeper move into national security AI. Cool!
🔵 Meta // Meta AI
Q3 Earnings: Meta reported $51.2B in revenue (up 26% YoY) but profit dropped sharply due to a $15.9B tax charge. The company also increased proposed AI spend up to $70–72B. Zuck still bullish on AI.
Stock Reaction: Shares fell ~7% after earnings as investors worried about Meta’s massive AI spending and delayed profitability despite strong ad growth. Seemingly not a concern for AI companies with better direction.
Reality Labs Loss: Meta’s XR division lost $4.4B in Q3, driven by high R&D and marketing costs for Quest and Ray-Ban Meta devices, which remain long-term bets. AI wearables will be a huge market. Not sure if glasses is the play.
🔴 xAI // Grok
Grokipedia Under Fire: xAI’s new encyclopedia was criticized for copying large sections from Wikipedia, raising questions about originality despite Creative Commons disclaimers.
Musk vs. Altman: Elon and Sam Altman reignited their feud over OpenAI’s for-profit shift, with Musk accusing Altman of betraying its mission and Altman calling Musk a self-interested rival. OpenAI was initially non-profit when Elon invested, but Elon also did leave the company. Both parties are somewhat right here and the back-and-forth accomplishes nothing.
Tesla Pay Vote Nears: Shareholders vote Nov 6 on Elon’s massive pay package tied to bold growth targets. Proxy firms oppose it, while ARK Invest and others back it. Seems more about control then the money. Worthwhile to follow to understand Elon’s time allocation across companies.
🤖 Impact Industries 👩🏫
Robotics // Home Automation
1X has unveiled NEO, a fully autonomous humanoid robot built for household chores like laundry, dishes, and tidying up. Priced at $20,000 or $499 per month, NEO combines advanced AI vision, motion control, and natural communication to perform everyday tasks safely around humans. The company, backed by OpenAI, says the robot will ship in 2026. May be somewhat tele-operated at first for some tasks. Will only get better with time. This is the future.
Education // AI Tutoring
'Super Teacher' unveiled an AI tutor for elementary schools designed to make one-on-one learning accessible nationwide. Founded by former Google product manager Tim Novikoff, the platform uses adaptive dialogue and real-time feedback to mirror the benefits of human tutoring at a fraction of the cost. Backed by leading education investors, Super Teacher is testing in U.S. classrooms and could mark a major shift toward personalized, affordable AI education. Great use case for AI.
🎙 Weekly Interview: 10 Minutes With Mike Alnakhaleh
Malik (Mike) Alnakhaleh
🏠 Background: Based in Columbus, OH, Mike A is a software engineer and educator with a B.S. in Computer Science (Summa Cum Laude) from Southern New Hampshire University. He’s passionate about using AI to make software more practical and accessible.
💼 Work: Mike is an AI/ML Developer at Cardinal Health, building enterprise apps powered by Gemini AI and RAG. He also consults on AI projects and previously taught Amazon’s backend tech stack.
🚀 Quote: “I would argue we've already achieved AGI personally, and what we're really trying to pursue is ASI.”
Condensed Interview — Mike Alnakhaleh
Noah Weisblat: Mike, do you wanna introduce yourself and tell us a little bit about what you do?
Mike Alnakhaleh: Yeah, of course. I’m currently working at Cardinal Health as a contract software developer, building web applications that integrate LLMs, image models, and classical ML systems. I also do AI consulting when it comes up, focusing on making AI more practical for real-world software.
Noah: What are your general thoughts on the AI space right now?
Mike: I think AI is both overhyped and underhyped, depending on who you talk to. There’s a lot of uncertainty — some people expect too much too fast, while others underestimate how close we actually are to massive breakthroughs.
Noah: How do you define AGI, and how close do you think we are?
Mike: Personally, I think we’ve already hit a form of AGI — systems that can do a variety of tasks at or above human level. What we’re chasing now is ASI, or artificial superintelligence. I think we’re only a few key breakthroughs away — things like continual learning, embodiment through robotics, and longer-term consistency.
Noah: What’s the biggest use case for AI in enterprise right now?
Mike: Retrieval-Augmented Generation (RAG), without question. Imagine an internal ChatGPT that can semantically search all your company’s documentation — HR files, technical docs, policies — and answer questions conversationally. It’s the most accessible, high-value way for enterprises to start integrating AI today.
Noah: When do you use a full AI agent versus a simple automation?
Mike: I’d use automation when you need consistent, quantitative results — like counting data points in a CSV. Agents make sense for qualitative tasks like summarizing or analyzing data, where there’s room for creativity or variability. For now, mission-critical systems should stay automated, but agents are ideal when you can afford flexibility.
👨💻 Practical Use Case: ChatGPT Personality
Difficulty: Basic
When GPT 5 came out, the personality and response style changed. There was a ton of backlash and while some prefer the quick, direct answers, others want deeper explanations or creative takes. With ChatGPT’s personalization settings, you can now customize how the model responds to you.
It’s pretty simple to try out for yourself:
👉 Click your name in the bottom left corner → Settings → Personalization
From there, you can:
Choose a personality (Default, Cynic, Robot, Listener, or Nerd)
Add traits like “Straight shooting,” “Encouraging,” or “Witty”
Set a nickname and describe your role (student, engineer, creator, etc.)
Once saved, your GPT adapts to your preferred tone, depth, and pacing. Whether you want faster summaries, more context, or extra humor, customization makes the AI feel a bit more in tune with someone you’d like to have a conversation with.
Here’s an example if you want a bit more context ⬇️
💻 Startup Spotlight
Extropic’s TSU Demo Case
Extropic: The computer chip that may redefine the future.
The Problem: Today’s CPUs and GPUs rely on binary ‘bits’ in the form of 1s and 0s, that rapidly flip back and forth to process data. These bits switch billions of times per second, consuming massive amounts of energy and compute power. A modern GPU contains around 5 billion bits, and all of that state-switching is a key bottleneck in scaling AI and advanced software systems.
The Solution: Extropic has introduced a new kind of chip: the Thermodynamic Sampling Unit (TSU). Instead of using deterministic bits, the TSU relies on probabilistic bits (p-bits) that use thermodynamic coupling to sample likely outcomes. This lets the chip predict whether a bit will be 1 or 0, rather than explicitly calculating it. The result is a radically different, more efficient computational model: one that could reduce compute requirements by up to 10,000x for certain tasks.
The Backstory: Extropic was founded by MIT PhD Josh Alman and has backing from top-tier investors like Kindred Ventures, HOF Capital, and NVIDIA. Their approach draws from both thermodynamics and statistical physics to build chips that "think" more like natural systems. It’s early days, but the implications are massive: if Extropic’s TSU works at scale, it could upend the entire foundation of digital computing and redefine what’s possible in AI, robotics, and simulation-heavy fields.
My Thoughts: I saw a lot of buzz about this last week upon launch but my foundational knowledge of computing needed a refresh, so I did some reading and asked chatGPT a bunch of questions. Now that I have a bit more of an understanding, I can say this has potential to be the single most important startup of the decade. It’s somewhere between regular computing and quantum computing, and it could allow for the compute that AI systems need in order to turn the corner into AGI or ASI. It’s still up in the air whether this actually works properly and at scale, but Extropic has taken a huge leap into the future with its demo and initial launch.
“It’s not likely you’ll lose a job to AI. You’re going to lose the job to somebody who uses AI”
- Jensen Huang | NVIDIA CEO
Anybody in the market for a home robot? Till Next Time,
Noah on AI
👋 Meet the Hoomans!
Chat with lifelike AI Humans who see, listen, and respond with real emotion. Whether it’s Maxine the witty Gen X’er or The Storyteller with a voice full of mystery, every conversation feels human.
💬 Try it now at Hoomans.chat where your AI talks like one of us.
