Boooooo

why is common sense not common at all?

[Boooooo @enderrith]

It's fascinating to watch someone stir my emotions and realize how little I actually care. I thought I would. But I don't. That's my defense mechanism — letting people close enough, but never all the way in.

[Google Research @GoogleResearch]

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: https://t.co/CDSQ8HpZoc

[Aakash Gupta @aakashgupta]

Thariq and the Claude Code team are shipping at a pace that’s becoming its own product strategy.

Slack integration Monday. Chrome Quick Mode yesterday. Auto-memory today. /insights, Fast mode, Figma sync, VSCode GA all in the last few weeks. A meaningful feature drop every 48 hours, and each one absorbs a capability that used to require a separate tool or vendor.

This is becoming a case study in how to build product in 2026. No big bang launches. No quarterly roadmaps gated by marketing cycles. Just relentless daily iteration where each feature compounds on the last. Memory makes /insights better. Slack context makes memory smarter. Fast mode makes the whole loop tighter.

The meta-lesson is hard to ignore. The teams shipping fastest right now are the ones treating product development the way their own AI tools work: continuous, contextual, compounding. Everyone else is still planning sprints while Claude Code is on its third feature this week.

Was playing with auto-memory this morning and updating the memory notes is addictive. Good luck keeping up with this team.

[Milk Road AI @MilkRoadAI]

Layoff Announcements:

1. US Government (DOGE): 317,000 employees
2. UPS: 78,000 employees
3. Amazon: 30,000 employees
4. Intel: 25,000 employees
5. Citigroup: 20,000 employees
6. Nissan: 20,000 employees
7. Nestlé: 16,000 employees
8. Microsoft: 15,000 employees
9. Bosch: 13,000 employees
10. Verizon: 13,000 employees
11. Dell: 12,000 employees
12. Accenture: 11,000 employees
13. Ford: 11,000 employees
14. Novo Nordisk: 9,000 employees
15. Procter & Gamble: 7,000 employees
16. HP Inc.: 6,000 employees
17. Heineken: 6,000 employees
18. Siemens: 5,600 employees
19. PwC: 5,600 employees
20. Dow Chemical: 4,500 employees
21. Salesforce: 4,000 employees
22. Lufthansa Group: 4,000 employees
23. ANZ Bank: 3,500 employees
24. GM (General Motors): 3,300 employees
25. ConocoPhillips: 3,000 employees
26. IBM: 2,700 employees
27. American Airlines: 2,700 employees
28. WiseTech: 2,000 employees
29. Morgan Stanley: 2,000 employees
30. Paramount: 2,000 employees
31. Starbucks: 2,000 employees
32. Target: 1,800 employees
33. Southwest Airlines: 1,750 employees
34. Meta: 1,500 employees
35. Applied Materials: 1,444 employees
36. Nike: 775 employees​
37. Kroger: 1,000 employees
38. eBay: 800 employees
39. Block Inc. (Square/Cash App): 1,100 employees

AI is officially replacing jobs at mass scale in the US.

Where will all of these people go?

[Boooooo @enderrith]

Hence this all-important principle: * to get more win-win outcomes one needs to negotiate with consideration given to what is most important to the other party and to oneself and know how to trade them.

[Boooooo @enderrith]

Complexity is usually a symptom of an unsolved upstream question.

[Ray Dalio @RayDalio]

https://t.co/tjmbT5ytUN

[Boris Cherny @bcherny]

I'm Boris and I created Claude Code. I wanted to quickly share a few tips for using Claude Code, sourced directly from the Claude Code team. The way the team uses Claude is different than how I use it. Remember: there is no one right way to use Claude Code -- everyones' setup is different. You should experiment to see what works for you!

[Min Choi @minchoi]

Less than 96 hours ago, Clawdbot (now Moltbot) broke the Internet.

Minds are blown. And people are already coming up with wild use cases.

8 examples:

[Min Choi @minchoi]

AI just fixed your life and business.

Claude Cowork is the $20/month AI employee everyone was afraid of 🤯

Train it once on your workflow + SOPs and it just runs... 24/7.

10 wild results (🧵👇)

1. Match up 1000s of day trades for tax season

[DAN KOE @thedankoe]

https://t.co/kfa0CkvdgU

[DAN KOE @thedankoe]

https://t.co/AiGjkp3513

[Min Choi @minchoi]

Anthropic literally dropped the Prompt Guide to 10x your Claude Opus/Sonnet 4.5

10 best practices you should know. Bookmark this. 👇

[Boooooo @enderrith]

Gained my super power: extrapolate at its peak

[Melvin Vivas @donvito]

45 use cases on how to use Claude

Practical examples across research, writing, coding, analysis, and everyday tasks

Article
https://t.co/rmd8uBVZ7Y

[Boooooo @enderrith]

@NoahEpstein_ Desktop

[Boooooo @enderrith]

@aaditsh AI

[Shubham Saboo @Saboo_Shubham_]

Claude Skills vs Claude Subagents vs Claude Projects clearly explained

[Alex Prompter @alex_prompter]

everyone's arguing about whether ChatGPT or Claude is "smarter."

nobody noticed Anthropic just dropped something that makes the model debate irrelevant.

it's called Skills. and it's the first AI feature that actually solves the problem everyone complains about:

"why do I have to explain the same thing to AI every single time?"

here's what's different:

- you know how you've explained your brand guidelines to ChatGPT 47 times?
- or how you keep telling it "structure reports like this" over and over?
- or how every new chat means re-uploading context and re-explaining your process?

Skills ends that cycle.

you teach Claude your workflow once.

it applies it automatically. everywhere. forever.

but the real story isn't memory. it's how this changes what's possible with AI at work.

[John Luke @yesknow]

@elonmusk Jensen Huang, CEO of Nvidia, provided some thought-provoking insights around 20 years ago.