Daily AI Newsletter: Here’s what the data says people ask ChatGPT

In today’s email:

👨🏻‍💻 Vibe coding has turned senior devs into ‘AI babysitters,’ but they say it’s worth it
👑 Gold is out-performing the Nasdaq 100 since the launch of ChatGPT
🔥 OpenAI announces new mentorship program for budding tech founders
🧰 8 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Here’s what the data says people ask ChatGPT

Key Takeaway: OpenAI’s new 62-page analysis of 1.1M+ chats shows ChatGPT is used mostly for personal “practical guidance” and writing help, with a young, increasingly female user base and fast-rising search-like queries.

More Insights:

Who’s using it: By June 2025, 52% of users had feminine first names; nearly half of chats came from 18–25-year-olds; growth is faster in poorer countries; OpenAI cites 700M+ weekly users.
Work vs. life: Non-work chats climbed to 73% (vs. ~50/50 in June 2024).
Top uses: “Practical guidance” is largest (28.3%), writing is second; “seeking information” grew steadily and reached #2 for consumers; coding-related chats were 4.2%.
Writing tasks: Editing/critiquing (38%) leads, then personal comms (28%), translation (16%), summaries/arguments (13%), fiction (5%).
Companionship is niche: Relationship/feelings advice is 1.9% of chats; “games/role-play” (incl. “AI girlfriend”) 0.4%, amid FTC scrutiny of companion bots; parental controls are in development.

Why it matters: If ChatGPT keeps shifting from workplace tool to everyday assistant, and substitutes for search, traffic, advertising models, and retail discovery could be reshaped, while regulators and designers race to balance massive consumer adoption with safety, transparency, and healthy usage.

Your Exclusive Invite for the World’s first 2-day AI Challenge (usually $895, but $0 today)

Join the online 2-Day LIVE AI Mastermind by Outskill- a hands-on bootcamp designed to make you an AI-powered professional in just 16 hours.

Usually $895, but for the next 48 hours, you can get in for completely FREE!

In just 16 hours & 5 sessions, you will:

✅ Learn the basics of LLMs and how they work.
✅ Master prompt engineering for precise AI outputs.
✅ Build custom GPT bots and AI agents that save you 20+ hours weekly.

🧠 Live sessions- Sat & Sun

Join in now (Limited free seats available)

Thinking Machines Lab wants to make AI models more consistent

Key Takeaway: Thinking Machines Lab says LLM randomness largely comes from how GPU kernels are orchestrated during inference, and claims tighter control there could make model outputs reproducible, boosting reliability and training.

More Insights:

$2B-backed lab led by ex-OpenAI CTO Mira Murati, with an all-star ex-OpenAI team.
Blog post (“Defeating Nondeterminism in LLM Inference”) by researcher Horace He kicks off the new “Connectionism” series.
Core claim: determinism is achievable by carefully managing GPU kernel stitching during inference.
Payoff: cleaner data and smoother reinforcement learning; more consistent answers for enterprises and scientists.
First product aimed at researchers/startups is due “in the coming months,” but details and ties to this work remain unclear.

Why it matters: If reproducible LLMs become real, AI shifts from “best-effort” to “verifiable”, unlocking auditability, compliance, and dependable research workflows, and moving the frontier from mere scale to systems engineering rigor.

OpenAI Ramps Up Robotics Work in Race Toward AGI

Key Takeaway: OpenAI is rebuilding a robotics team, especially around humanoids, hiring specialists and leaning on teleoperation and simulation to push embodied AI toward AGI.

More Insights:

New hires (e.g., Stanford’s Chengshu Li) signal a focus on humanoid-capable household skills and evaluation benchmarks.
Job posts call for teleoperation, Nvidia Isaac simulation, and a mechanical engineer with “1M+” manufacturing experience, hinting at scalable hardware ambitions.
Whether OpenAI builds robots, partners, or uses off-the-shelf hardware remains open, but the stated mission is “general-purpose robotics” toward AGI.
Competition is fierce (Figure, Agility, Apptronik, Tesla, Google), and experts like Brown’s Stefanie Tellex see no inherent OpenAI edge yet.
Context: OpenAI paused robotics in 2021 after notable work (Rubik’s Cube hand), restarted in 2024; humanoid funding has surged ($5B+ since 2024, with a $5T market by 2050 forecasts).

Why it matters: If large language models are hitting diminishing returns, grounding intelligence in perception and action could be the next leap, whoever masters the data flywheel of teleoperation → simulation → real-world robots may define the future of both AI capabilities and how physical work gets done.