Sam Altman’s Post-GPT-5 Playbook: Hardware, Chrome(?), and Trillions for AI

Key Takeaway: Over an unusually candid dinner, Sam Altman framed GPT-5’s bumpy rollout as a speed bump on OpenAI’s bigger push into consumer apps, devices, and massive infrastructure, while promising warmer models, clearer transitions, and a neutral product stance.

More Insights:

• GPT-5 backlash triggered a quick course-correction: GPT-4o and the model picker returned, and GPT-5 is being tuned to feel “warmer.”

• Despite criticism, usage surged: API traffic doubled within 48 hours; ChatGPT hit new daily highs and now reaches ~700M weekly users.

• New fronts beyond ChatGPT: incoming “applications” CEO Fidji Simo to oversee multiple consumer apps (AI browser, social), plus interest in buying Chrome if forced to sell.

• Big bets on interfaces and compute: funding a Neuralink rival, a Jony Ive-designed device, and plans to spend trillions on data centers due to GPU scarcity.

• Guardrails and stance: limit parasocial harms, avoid “ideological” defaults, but allow user-steerable tone/perspective; promise better model-deprecation transition periods.

Why it matters: This signals a shift from “model as product” to “AI as platform” backed by colossal infrastructure, and whoever marries reliable behavior, humane UX, and distribution (apps, devices, browsers) will shape how billions interact with the web and, increasingly, with each other.

The 16-hour AI challenge: Learn AI, build Products and Earn $10K/Month

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, 10 AM - 7 PM.

In 16 hours & 5 sessions, you learn:
✅ Understand  LLMs fast
✅ Master prompts for Precise AI results
✅ Build bots to save 20+ hrs/week

$5100+ worth bonus: Day 1: 3000+ Prompt Bible, Day 2:Roadmap to make $10K/month with AI, additional bonus: Your Personal AI Toolkit Builder

Grammarly says its AI agent can predict your grade

Key Takeaway: Grammarly launched nine specialized AI agents, free for Free and Pro users inside its new “Docs”, spanning grading predictions, proofreading, paraphrasing, citations, and plagiarism/AI detection to guide students and assist educators.

More Insights:

• AI Grader tailors feedback using uploaded course details and publicly available info about the instructor, then estimates a likely grade.

• Reader Reactions predicts questions readers might have; Proofreader adds inline suggestions; Paraphrase adapts tone, audience, and style.

• Citation Finder auto-generates properly formatted references; Expert Review gives topic-specific feedback.

• Plagiarism Checker scans large databases and the web; AI Detector scores the likelihood text is human- vs AI-written (both Pro-only at launch).

• Available now at no extra cost for Free/Pro; Enterprise and Education support is promised later this year with more agents to come.

Why it matters: This signals a shift from one-size-fits-all chatbots to targeted academic tools, promising faster, more personalized learning while raising fresh questions about grading fairness, privacy (using “public info” about instructors), and what AI literacy should mean in modern classrooms.

Gamblers Now Bet on AI Models Like Racehorses

Key Takeaway: Prediction markets like Kalshi (and offshore sites) now let bettors wager on which AI model “wins,” turning social-media signals and leaderboard whispers into tradable odds where the most online, best-informed traders pocket the profits.

More Insights:

• Trading volume in AI markets hit about $20M this month; Kalshi’s AI activity is ~10× higher than early 2025.

• Example play: a trader bought Gemini at ~$0.40 and sold at $0.87 on “Best AI This Month,” netting ~$10K before settlement.

• Markets span model rankings and industry events (e.g., “Best AI at year-end,” “US federal AI regulation passes,” “Sam Altman gets equity in OpenAI”).

• Edge comes from X/Discord chatter, LMArena blind tests, and even scraping product back-ends for telltale updates.

• Economists note rising volume rewards information advantage, squeezing casuals as “sharks” arbitrage across platforms.

Why it matters: This financializes AI progress itself, creating incentives to mine rumors, exploit leaks, and move sentiment, blurring lines between tech development, market speculation, and policy signals, and raising fresh questions about fairness, manipulation, and regulation.

Other stuff

• What My Daughter Told ChatGPT Before She Took Her Life

• Class-action suit claims Otter AI secretly records private work conversations

• Blue-collar jobs are gaining popularity as AI threatens office work

• OpenAI Staffers to Sell $6 Billion in Stock to SoftBank, Other Investors

• AI experts return from China stunned: The U.S. grid is so weak, the race may already be over

• Meta’s flirty AI chatbot invited a retiree to New York.

Hope you enjoyed today's newsletter

⚡️ Join over 300,000 people using the Superpower ChatGPT extension on Chrome and Firefox.