@elonmusk@AnthropicAI@SpaceX If there is not enough compute for Anthropic to keep up and supposedly unlimited token demand, why is Anthropic not using all the capacity available ?
@alexeheath Tokenmaxxing…What am I missing here? So all companies are budget constrained and there is the potential for complete tokenmaxxing cost abuse especially from employees who don’t know what they’re doing. How can that lead to anything but large cost and low productivity.
Tokenmaxxing…What am I missing here? So all companies are budget constrained and there is the potential for complete tokenmaxxing cost abuse especially from employees who don’t know what they’re doing. How can that lead to anything but large cost and low productivity.
@sama What am I missing here? So all companies are budget constrained and there is the potential for complete tokenmaxxing cost abuse especially from employees who don’t know what they’re doing. How can that lead to anything but large cost and low productivity.
AI in 2026:
Context = Memory.
Give AI your context, your ideas, your notes,
your memory.
• Obsidan — capture it all
– no thought too small
– no plan too big
– fleeting ideas → plans → strategy → decisions
• OpenClaw — agents process your obsidian notes
Obsidian becomes your voice.
OpenClaw becomes your AI leverage.
Capture it now.
Thank me, L8rs.
DM's are open for consulting queries - if you understand that you must make a move, but it's moving too fast.
Frontier models are exceptionally efficient, intelligent, and useful. For agents, context is now the bottleneck.
Enter the context layer, which bridges the gap from an enterprise's messy data to actionable context, packaged for agents.
We're seeing three distinct verticals emerge in the context layer space:
- Data gravity platforms
- Existing AI data analysts
- New, dedicated context layer companies
Read the full piece by @JasonSCui and @JenniferHli: https://t.co/ftyF4lYIFK
My information consumption is now 1/4 X, 1/4 podcast interviews of the smartest practitioners, 1/4 talking to the leading AI models, and 1/4 reading old books. The opportunity cost of anything else is far too high, and rising daily.
My information consumption is now 1/4 X, 1/4 podcast interviews of the smartest practitioners, 1/4 talking to the leading AI models, and 1/4 reading old books. The opportunity cost of anything else is far too high, and rising daily.
7/7
This is not software in the old sense.
It is a living support system for kids and parents.
A new AI universe with payments built in.
Register for early access: https://t.co/eW0sI0sh6g
1/7
The world is dead. Long live the world.
20 years ago, I launched the world’s first mobile App Store, distributed software globally, and led it to an IPO.
A decade later, Apple launched the App Store, and software started to eat the world.
6/7
Kids do not need more forms, menus, and software.
They ask their PAL.
Behind the scenes, swarms of invisible AI assistants get to work through a specialist support crew of BrainPal agents.
That is the shift.
5/7
BrainPal is building a kids-and-parents OS powered by specialist AI companions, or PALs.
What matters is not one clever bot.
What matters is a swarm of AI agents working together in real time.
4/7
That is what OpenClaw unlocked for us at BrainPal.
We can now build at a fraction of the cost, in a fraction of the time, with far more power.
So much so that I refused large VC funding and oversized equity demands.
We walked. As for https://t.co/IETCaiQezX - jokers!
7/7 That is a digital workforce acting at the moment of decision. What used to take weeks of product design and engineering can now happen behind the scenes in seconds.
We are building now. If you want help doing the same: https://t.co/vZ7i2sr7vz
6/7
Example:
A child goes to buy a Coca-Cola.
MoneyPAL scans the shelf.
HealthPAL flags it.
ParentPAL blocks it.
RewardPAL offers an incentive for water.
MoneyPAL completes payment.
Parent gets updated.
That is not a chatbot.