🌟 Editor's Note: Recapping the AI landscape from 07/28/25 - 08/04/25.
🎇✅ Welcoming Thoughts
Welcome to the 4th edition of NoahonAI.
What’s included: company moves, a weekly winner, AI industry impacts, practical use cases, and more.
Consulting: I’ve spoken with a few folks regarding AI workflow consulting for businesses. I’ll be launching that soon.
We’re right on the cusp of the 2nd phase of AI (Agentic).
Still patiently waiting on GPT-5.
A reminder for when you see headlines related to recalled updates or faulty releases: Don’t panic, we’re still very early.
Another great interview for startup founders this week.
Will be releasing a blog in the next week or so expanding on this week’s practical use case: GPT Wrappers.
I feel like NVIDIA either has a slow week or a monumental one, with little in between. Given their position, neither one is a bad thing.
Let’s get started—plenty to cover this week.
👑 This Week’s Winner: Anthropic // Claude
1st time winner Anthropic! It appears a lot of their long-term bets are paying off. Their biggest win is a come-from-behind-lead over OpenAI in market share for enterprise users. Anthropic reportedly now controls 32% of the enterprise AI market compared to OpenAI’s 25%.
In other news, Anthropic has made a power play to cut off OpenAI’s usage of Claude entirely. Their justification is around OpenAI technical staff using the product ahead of GPT-5.
Claude Code has announced usage limits crackdowns which can be framed as positive or negative, but if you’ve been following their drama surrounding random usage cutoffs and financial questions, I appreciate the stability of a firm announcement on user limits. It also indicates high user interest in the product.
Speaking of high interest, Anthropic hasn’t had much trouble in securing lead investor Iconiq Capital, for a ~ $5B funding round at what appears to be a $170B valuation 😳.
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
Just Insane Numbers: OpenAI hit a $12B Annualized revenue rate, doubling revenue in the first seven months of the year, reached 700M weekly active users, and closed an $8.3B funding round at a $300B valuation.
Study Mode Launch: OpenAI just added “Study Mode” to ChatGPT. It’s designed to guide students through interactive questioning instead of direct answers. AI is outstanding for learning if used right.
Privacy Concerns: OpenAI had to remove a new feature after it resulted in GPT conversations being indexed by Google. This in addition to last week where Sam Altman revealed GPT conversations were not impervious to legal warrants.
🟣 Google // Gemini
Deep Think Rollout: Google released Gemini 2.5 Deep Think to AI Ultra subscribers ($249/mo). This is the model Gemini used for the IMO and its reasoning capabilities are off the charts. It scored a leading 34.7% on Humanity’s Last Exam.
Gemini AI Impact: Google consolidated ~200 engineers into a core Gemini team to unify AI across Search, Gmail, Docs, Chrome, and Android, with a mid-tier “AI Lite” subscription under discussion.
Security Fix After Gemini Vulnerability: Google patched a critical security flaw in the Gemini CLI tool that allowed hidden code execution in README-style files.
🔵 Meta // Meta AI
Earnings Call → Stock Surge: Meta’s Q2 revenue has increased 22% YoY to $47.5B. Shares jumped 12% post-earnings.
Zuck Eyes ASI: Mark Zuckerberg said Meta is nearing personal superintelligence. “Over the last few months we have begun to see glimpses of our AI systems improving themselves”.
Meta Offers $250M Deal to AI Wunderkind: Meta keeps spending on talent as they hire 24‑year‑old AI researcher Matt Deitke, giving him a $250M compensation package.
🔴 xAI // Grok
Launch of Grok Imagine: xAI just rolled out Grok Imagine. Elon has dubbed it “AI Vine” and it is notable for super fast image and video creation. Good: speed, remixing videos, still image → video. Poor: No actual voice - just sound effects, quality (esp. in comparison to Veo3).
Internal Grok Leaks: Internal project “Skippy” asked workers to record themselves in an effort to help Grok recognize emotion (Just use VERN!). Leaked prompts also revealed how Grok integrates into Tesla vehicles for in‑car voice use.
Tesla Doubles Down on Elon: xAI adjacent - Tesla awarded Elon nearly $30B worth of stock in an effort to keep him on their management team. This is important for those looking at Elon to focus on Tesla and humanoid robotics versus other ventures.
⚪️ NVIDIA
China Probes NVIDIA: China’s Cyberspace Administration summoned NVIDIA executives over alleged backdoors in its H20 AI chips (NVIDIA denies). Notable given this is something the US Govt. has discussed implementing.
NVIDIA Isn’t Just Hardware: NVIDIA updated ForecastNet 3, its latest deep-learning model for rapid, global multi-day weather predictions. Currently boasting 15-day forecasts in less than a minute.
Earnings Call Q2 Set: NVIDIA scheduled its Q2 earnings call for August 27th at 5pm EST. CFO Colette Kress will be present to comment on results.
🔬 Impact Industries 🏢
🔬 Science // Autonomous AI
A new preprint describes hierarchical AI systems capable of auto-generating hypotheses, running experiments, and even spawning specialized sub-agents in real time. Researchers Ling Yue, Shimin Di, and Shaowu Pan argue this architecture could compress decades of scientific progress into days by letting AI self-organize entire research pipelines. If scalable, it’s a blueprint for billions of AI “scientists” working in parallel.
🏢 Enterprise // Mandatory AI Adoption
Yahoo Japan is going all-in on AI at the organizational level. The company just mandated daily AI use for all 11,000 employees with the goal of doubling productivity by 2028. Alongside required training, usage tracking, and company-wide integration, it’s one of the boldest corporate bets on AI yet, and a glimpse of what a full-scale, enforced rollout looks like inside a major business.
👨💻 Practical Use Case: GPT Wrappers
Difficulty: Mid-Level
A GPT wrapper is a mini tool or app you build around ChatGPT. Instead of typing into ChatGPT manually, you write code that sends the prompt for you, often with context, logic, and structured outputs built in. This turns GPT from a general assistant into a focused tool that does one job really well. A lot of early AI startups were simply GPT wrappers made to look pretty.
GPT wrappers are best for specialized, repetitive tasks. Such as:
Summarizing team-specific documents and sends highlights to Slack
Generating social captions from a specific product catalog
Optimizing landing page copy based on conversion data
Imagine you’re running a Facebook Ad campaign and are getting lots of traffic, but visitors are not buying. You can build a GPT wrapper that takes in web visits and conversion rate data, and returns updated HTML to your site. Every 100 visits, the AI will receive conversion results and decide what copy to send back based on the recent tests.
Once the wrappers are built, you can connect them to other tools and you will have a Hybrid Workflow. Since you’re building on Chat, and not creating your own LLM, the build difficulty : tangible results ratio is pretty good.
I’ll be putting together a blog to show you step-by-step how to make your first GPT wrapper, in the meantime, feel free to check out the video below or ask Chat for help getting started!
🎙 Weekly Interview: Manny Larcher
Manny Larcher
🏠 Background: Based in Columbus, OH, Manny exited a venture-backed startup in 2017. More recently, he was a Venture Analyst, advising early-stage startups across tech and healthcare.
💼 Work: Manny is CEO of Colaeb for business growth and Co-Founder of Stopwatch Creative. He’s a startup advisor, mentor, consultant, investor, and writer.
🚀 Quote: “You should just execute and not overthink. If there was ever a time to not overthink, this might be it.”
Condensed Interview Transcript – Manny Larcher
Noah Weisblat:
I’m here with Manny Larcher. Manny, tell us about yourself.
Manny Larcher:
Hey Noah. I’ve got over a decade of experience in tech, mostly at the intersection of tech and marketing. I’ve been part of venture-backed teams, run a software company and a consulting agency, and worked in venture capital and angel investing. I’ve worn a few hats and I’m excited about where things are going.
Noah:
We connected a couple of years ago when I was working on Klout Machine and you were at JumpStart. What are you focused on now?
Manny:
Before JumpStart, I was running a company called Collaborate and Elevate — now known as Collaeb. It’s a platform that connects businesses to new customers, capital, talent, and advertising solutions. We incubated Collaeb out of Stopwatch Creative, which I started in 2017. JumpStart had just raised a $50M fund, and I stayed through the duration of that before shifting back full-time to Stopwatch and Collaeb earlier this year.
Noah:
Let’s talk AI. What’s your general take on the space?
Manny:
It’s exciting and extremely useful. I’ve been using AI early on for both personal efficiency and customer experience. There are companies I’ve invested in, like Taja.ai, built on ChatGPT, and they’re doing really well. We also use it internally to streamline operations and boost customer outcomes.
Noah:
Which tools are you using the most?
Manny:
ChatGPT, for sure. I use others, but ChatGPT is at the core of a lot of what I do. Taja is also a go-to — especially for content editing and repurposing video.
Noah:
How are you using those tools at Collaeb?
Manny:
With Taja, I take long-form videos like this one, upload them, and it outputs short-form clips I can post across social channels. That saves time and increases engagement. In July, I posted more and saw direct results — more followers, engagement, and customers. Tools like this let me and my team move faster and work smarter.
Noah:
What’s your take on other players in the space — the “NVIDIA 5” like OpenAI, Gemini, xAI, Perplexity, etc.?
Manny:
They’re all interesting. I’ve been diving deeper into NVIDIA recently. Whether or not you’re building directly on their platforms, there are ways to participate — even through investing. It’s smart to watch companies like ARM, TSMC, and others shaping the AI infrastructure.
Noah:
You mentioned angel investing. When it comes to evaluating AI startups, how do you do it? And do you use AI to help with that process?
Manny:
Yes, I absolutely use AI to help evaluate companies. One tip I give founders: copy your website and paste it into ChatGPT, then prompt it to evaluate from a VC perspective. It’ll give you a preview of how investors might see your pitch. As for evaluating companies, traction is everything. Taja, for example, has 5,000+ customers with a tiny team — that tells me they’re solving a real problem.
Noah:
What advice would you give to a college student or recent grad looking to use AI to start or grow something?
Manny:
Start. Don’t wait for perfection. Too many people hesitate because they’re afraid of what others might think. Build something that provides value. If someone pays you and is happy, that’s all that matters. Execution beats overthinking every time.
Noah:
Some people think AI is a bubble — that it’s useful but overhyped. What would you say to them?
Manny:
I’d say they’re wrong. AI is making entrepreneurship more accessible. Before AI, 9 out of 10 small businesses failed. Now we’ve got tools to help founders operate leaner and smarter. Execution is what separates successful founders — not complaining or waiting.
Noah:
From your perspective in venture and marketing, what industries do you think AI will disrupt most?
Manny:
Any industry with “comfortable” roles that aren’t measured by output — marketing, legal, even internal operations. AI is making it easier to track productivity. At the same time, the best talent is getting rewarded at a level we’ve never seen — $200M+ compensation packages. It’s going to widen the gap between high-performers and coasters.
Noah:
How much more efficient are you now compared to, say, 2022?
Manny:
At least 50x. That’s not an exaggeration. AI tools have changed everything about how I work and what I’m able to get done.
Noah:
Do you use AI for anything personal?
Manny:
I’ve definitely asked it what to wear before a date. It’s scary accurate, especially when it knows your content and context. But I try not to rely on it for everything.
Noah:
Where do you see AI in 5–10 years?
Manny:
Honestly, I don’t think in those timelines. Everything’s moving so fast that what we expect in 5 years might happen in 6 months. I focus on what I can do in the next 6–12 months — that’s where the opportunity is.
Noah:
Anything else you want to plug before we wrap?
Manny:
Yeah — check out collaeb.com. We’re running a free, no-equity accelerator. You can apply on the site. I’m also on LinkedIn and Instagram — just search Manny Larcher. And most importantly: execute. Don’t overthink.
Noah:
Appreciate your time, man. Great stuff.
📲 Startup Spotlight
Meatball Social
Meatball Social: A WhatsApp assistant to automate your life.
The Problem: Existing WhatsApp AI assistants don't work in groups, or act proactively.
The Solution: Our AI assistant lives in WhatsApp and automates your day: from smart daily briefings to to-do summaries, workout accountability, lead scouting alerts, cross-platform search, second-brain organization, and even group chat entertainment with roleplay characters. It’s your productivity engine and social sidekick in one.
The Market: The smart assistant market is booming, with millions already using AI-driven personal tools. Valued at $2.23 billion in 2024, it's projected to surge to $56.3 billion by 2034, reflecting growing demand for everyday automation and intelligent support.
Why Now? AI is evolving from reactive tools to proactive coaches, flagging and handling what we forget or miss, without needing to be asked.
The Backstory: “I left my PM role where I led a product team from $0-13M to focus full-time on this. We're building tools that help people become the best version of themselves. Right now, we're looking for early users to help us shape the product through tight, hands-on iteration over the coming months.”
Contact: [email protected]
“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
I’m old enough to remember it being headline news when a company raised $8B. Till next time,
Noah on AI
📖 Recommended Next Read
Marketing Junto
Marketing Junto is a weekly newsletter about the ever-changing landscape of digital marketing. From web design trends to SEO & paid search to social media, it’s all covered.
