Michael Lee

1.9K posts

Michael Lee banner
Michael Lee

Michael Lee

@_leebme

Author, Ex big tech, Londoner, Devoted Family man and entrepreneur. Write about AI realities at work. ‘Dadager’ to immensely talented son William.

London, England Katılım Temmuz 2008
723 Takip Edilen462 Takipçiler
Michael Lee
Michael Lee@_leebme·
@markgurman This could have been a killer update. Less weight, stronger ‘canvass/canopy’ (whateverthe hell its caused), foldable, non-bra case - AirPods Max Air Pros. Awesome
English
1
0
6
1.5K
Mark Gurman
Mark Gurman@markgurman·
Power On: Apple’s AirPods Max 2 is a prime example of a blurring line between marketing and actual innovation. My take on why giving the new model a “2” brand is a stretch. bloomberg.com/news/newslette…
English
50
32
648
156.8K
Michael Lee
Michael Lee@_leebme·
My son William’s (15) amazing composition and performance of ‘Prelude 2’. Self taught and plays by ear. Will is neurodiverse but god has given him an incredible gift. More to come ..
English
0
0
0
17
Classical Music: Our Heritage
Listening to classical music is an act of resistance. Against the algorithm. Against the 30-second attention span. Against the idea that everything must be easy and instant. Choose difficulty. Choose depth. Choose beauty.
English
88
222
1.5K
26.4K
Michael Lee
Michael Lee@_leebme·
@ihtesham2005 This is great. I added this as a skill but used part 2 as an 'opt-in' thinking pattern (Feynman) and make it ask at each stage / infer its business /learning before musk ending. It was crazy detailed on LLM training. I learnt a lot! thanks. I'd share the .MD I made if I could..
English
0
0
1
36
Ihtesham Ali
Ihtesham Ali@ihtesham2005·
🚨BREAKING: Claude has a secret mode called "First Principles Breakdown." It strips any complex topic down to its raw fundamentals like Elon Musk thinks through problems. Here's how to activate it:
Ihtesham Ali tweet media
English
59
494
4K
454K
Michael Lee
Michael Lee@_leebme·
Claude Skills. They are becoming like the 'collectibles' in video games.. Just me? Er ok then.
English
0
0
0
11
Michael Lee
Michael Lee@_leebme·
The AI slop on here now is out of control. It’s beyond the joke. Can’t anyone just write a genuine post anymore using their god given brain. We should be all be muting this crap from our timelines.
English
0
0
0
11
Michael Lee
Michael Lee@_leebme·
@heygurisingh Add “here’s what/why nobody tells you / knows / etc “ another ringer.
English
0
0
0
278
Guri Singh
Guri Singh@heygurisingh·
Holy shit... someone just open-sourced the cheat code for making AI writing undetectable. It's called stop-slop and it strips every known AI tell from your prose automatically. No rewriting tools. No paraphrasers. No "humanizer" apps. Here's how it works: → A single SKILL.md file you drop into Claude Code, Cursor, or any system prompt → Bans 50+ AI filler phrases your readers are already tired of → Kills structural clichés like dramatic fragmentation and binary contrasts → Forces sentence rhythm variation so your writing doesn't sound robotic → Scores your draft on a 50-point scale across 5 dimensions The wildest part? It's not a tool. It's not a SaaS product. It's a markdown file with rules. That's it. And it works better than any $29/month "humanizer" on the market. One file changes how your AI writes everything. 809 GitHub stars. MIT licensed. 100% Open Source. (Link in the comments)
Guri Singh tweet media
English
52
89
869
88.2K
Nebraskangooner
Nebraskangooner@Nebraskangooner·
One thing I need to work on is getting in late on trends because I have no idea how I would enter on a chart like this even though I think it goes higher. Anyone have any good tips? @Truecrypto @BigCheds @PeterLBrandt
Nebraskangooner tweet media
English
21
7
126
27.9K
Michael Lee retweetledi
Josh Kale
Josh Kale@JoshKale·
FINALLY! Anthropic just shipped the missing piece between OpenClaw and Claude Cowork… Remote control You can now text Claude from your phone using “Dispatch” and it: - Executes tasks on your desktop. - Reads your local files. - Searches your Slack and email. - Builds presentations. - Organizes your folders. One persistent conversation. You send it a mission from your pocket. You come back to finished work sitting in your file system. Your phone just became a remote control for an AI employee that lives on your computer.
Felix Rieseberg@felixrieseberg

We're shipping a new feature in Claude Cowork as a research preview that I'm excited about: Dispatch! One persistent conversation with Claude that runs on your computer. Message it from your phone. Come back to finished work. To try it out, download Claude Desktop, then pair your phone.

English
86
133
1.7K
364.2K
Michael Lee
Michael Lee@_leebme·
Is it me or is Claude being dumb lately on 4.6? For non coding related work it’s become way to agreeable again and making mistakes (ChatGPT vibes)🤦‍♂️
English
0
0
0
27
Michael Lee
Michael Lee@_leebme·
I was laid off due to tech and AI (ML called then) implementions across our business model. Coming from Digital Ad tech sector. I managed to switch to customer side/consult and find the pain problems to which eventually I built businesses around. Not easy task - owner mindset needed now.
English
0
0
0
601
Andy
Andy@PositivFuturist·
All my friends in high skilled jobs burning out and stressing about redundancy.. rough out there right now.
English
37
34
794
25K
Michael Lee
Michael Lee@_leebme·
@LloydLegalist Op- the undisputed GenX don of 80s. My parents took us to Florida in the 80s and coming from the UK it was like a wonderland then. I turned into Op man on holiday. Loved it. That surf vibe then was so fresh.
English
0
1
5
296
Lloyd Legalist
Lloyd Legalist@LloydLegalist·
Props to my fellow Gen Xers who rocked a velcro wallet and can still hear that loud rip every time it was opened. #GenX
Lloyd Legalist tweet media
English
117
148
3.2K
66.9K
parker
parker@parkerhendo·
I would pay 5-10x more for claude max if regular claude, cowork and claude code were all combined and could share context with each other. having things split across all these makes me sad.
English
109
30
1.3K
71.5K
Tansu Yegen
Tansu Yegen@TansuYegen·
Life for a girl dad can feel this fast sometimes🤍
English
88
1.4K
7.1K
759.1K
Tom Goodwin
Tom Goodwin@tomfgoodwin·
It seems a lot like AI is extremely bad for almost everyone’s mental health.
English
129
64
702
17K
Michael Lee
Michael Lee@_leebme·
Top tip, ask Claude to create a puns.md file. during your co work-day have a bit of fun with random gags from your AI partner. You'll never beat it!
English
0
0
0
23
Michael Lee
Michael Lee@_leebme·
Interesting. Cutting through the hype of this repo and it’s a prompt library, albeit an extensive one. Not a full stack agent list. Still very useful though.
Aakash Gupta@aakashgupta

This repo has 10k stars and almost none of them know what they actually starred. Each “agent” is a markdown file you copy into ~/.claude/agents/ or .cursor/rules/. There’s no coordination layer, no task handoff, no shared memory between agents. You manually invoke one persona at a time inside your existing coding tool and it responds with that personality’s specialty baked in. That’s useful. Genuinely useful. Having a “Security Engineer” persona that defaults to threat modeling when you ask it to review code saves you from writing that system prompt yourself. Same with a “UX Researcher” persona that structures feedback around usability heuristics instead of generic suggestions. But you are the orchestration layer. You decide which persona to activate, what context to pass between them, and how their outputs connect. The 10k stars in 7 days tells you something real about demand. Developers want specialized AI teammates, and they’ll star a repo that even approximates that vision. The gap between “I want an AI frontend developer, backend architect, and DevOps engineer collaborating on my project” and “I can load different system prompts one at a time” is where the actual hard problem lives. Multi-agent coordination, shared project state, autonomous task decomposition, conflict resolution when the Backend Architect’s schema breaks the Frontend Developer’s component structure. That’s the engineering problem nobody has solved cleanly yet. The repo is a well-organized starting point for prompt engineering. The viral framing as an “AI agency” is the packaging. And the mass enthusiasm tells you exactly how much latent demand exists for the real thing when someone finally builds it.

English
0
0
0
48