JANUARY 22, 2026

TODAY: ​Anthropic, OpenAI, and Google are now making a serious play for teachers and students. OpenAI may release a supposed AirPod Killer in 2026, although I'm not sure Apple should be panicking any time soon.

That, plus more coverage from Davos, good news about the AI gains starting to show in the real economy, and a slew of new prompts to improve outputs. Last but not least, a DIY algorithm creator that gives you more control over what you see and interact with online.

• PromptCraft: Prompts and Discussions to Avoid AI 'Tells'; Fiction Prompts; and 12 Constraint-Driven Tactics for Getting Better Outputs
• THE HEADLINES: ​How Anthropic, OpenAI, Google are trying to win over teachers; OpenAI may release supposed AirPod Killer in 2026; More from Davos; AI gains start showing in the economy​
• Tool Spotlight: Trace: Design Your Own Algorithm to Replace Doomscrolling

TRY THESE! Prompts and Discussions to Avoid AI 'Tells'; Fiction Prompts; and 12 Constraint-Driven Tactics for Getting Better Outputs

• ​Recognising ChatGPT-Written Text And Fixing It: OP explains why AI-written content feels generic: we omit crucial context like purpose, audience, location, constraints, and tone. They list missing dimensions for a sample business plan prompt and reference aichat.guide, which asks structured questions to build richer prompts.
• ​Prompting ChatGPT As A Fiction Co-Writer: 59-year-old author wants to translate novels into English and get unstuck on story beats using ChatGPT. Thread centers on prompt ideas for style-preserving translation, line editing, brainstorming branches, and keeping AI suggestions aligned with an existing voice.
• ​Twelve Constraint-Driven Prompts For Better Output: OP shares 12 reusable patterns (Reverse Brief, Scorched Earth Review, Competitor Autopsy, Stupid Questions Protocol, Priority Interrogation, Historical Replay, Confusion Detector, Five-Year-Old Test, Constraint Generator, Objection Forecast, Template Extraction, Energy Mapper) that force specificity instead of vague “help me” prompts.

HIGHLIGHTS: ​How Anthropic, OpenAI, Google are trying to win over teachers; OpenAI may release supposed AirPod Killer in 2026; More from Davos; AI gains start showing in the economy​

1. ​Prosus CEO: Don’t underestimate China’s AI capabilities​
2. ​Harari Warns: AI to Trigger Global Identity, Immigration Crises by 2026​
3. ​AI company Eightfold sued for helping companies secretly score job seekers​
4. ​All anyone wants to talk about at Davos is AI and Donald Trump​
5. ​Cisco CEO Chuck Robbins on building infrastructure for the AI era​
6. ​Michael Burry Warns Government Intervention Won’t Stop AI Bubble Burst: ‘The Problem Is Too Big To Save’​
7. ​Anthropic, Google and Microsoft fight to win teachers​
8. ​YouTube chief says ‘managing AI slop’ is a priority for 2026​
9. ​Medea: An omics AI agent for therapeutic discovery​
10. ​Amazon One Medical introduces agentic Health AI assistant for simpler, personalized, and more actionable health care​
11. ​Survey: Faculty Say AI Is Impactful—but Not In a Good Way​
12. ​NFL legend Jimmy Johnson slams AI video circulating across social media​
13. ​Introducing AI Glasses Impact Grants to Advance Wearable Technology for Good​
14. ​Corporate AI adoption could take ‘even longer’ than 10 to 15 years, says Thoma Bravo founder​
15. ​How to really spot AI-generated images, with Google’s help​
16. ​“Largest Infrastructure Buildout In Human History”: Jensen Huang on AI’s “Five-Layer Cake” at Davos​
17. ​How College Faculty Assess the Present and Future of Higher Education in the Age of AI​
18. ​Mastercard moves to set the rules for AI commerce​
19. ​Gates and OpenAI team up for AI health push in African countries​
20. ​Palantir CEO says AI ‘will destroy’ humanities jobs but there will be ‘more than enough jobs’ for people with vocational training​
21. ​Salesforce CEO Marc Benioff to governments across the world: Don’t make these ‘suicide mistake’ everyone did with Facebook, Instagram and other social media platforms​
22. ​Building a community-led future for AI in film with Sundance Institute​
23. ​New AI-powered video editing tools in Premiere, plus major motion design upgrades in After Effects​
24. ​Palantir CEO suggests AI ‘bolsters civil liberties,’ warns Europe falling behind US and China​
25. ​These iPhone AI apps expose your data, and they’re all over the App Store​
26. ​Cisco and OpenAI redefine enterprise engineering with AI agents​
27. ​OpenAI’s Sweetpea: ChatGPT-Powered Earbuds to Rival AirPods in 2026​
28. ​Elon Musk says Tesla’s restarted Dojo3 will be for ‘space-based AI compute’​
29. ​DeepMind CEO Says Chinese AI Firms Are 6 Months Behind the West​
30. ​VoidLink: Evidence That the Era of Advanced AI-Generated Malware Has Begun​
31. ​AI Gains Starting to Show in the Real Economy​
32. ​Gemini in Chrome is getting “Skills” as it moves toward becoming a full AI agent​

Trace: Design Your Own Algorithm to Replace Doomscrolling

Trace is a personalized “signal feed” that pulls content from multiple places—like Hacker News, Reddit, X/Twitter, YouTube, arXiv, and blogs—then delivers a tighter daily digest instead of an endless scroll.

It’s built to learn from your interactions over time (what you open, skip, and save), so the feed gets more relevant without you constantly tuning filters and lists. If you like staying informed but hate attention-grabbing algorithms, Trace is trying to give you a calmer, more intentional alternative.

Learn More: Trace​

​Read All mAIn Street Back Issues Here​

Your input fuels the content. Let me know how I did today!