Building https://nitter.cf/t.co/6qrAMdfhtG @ClawBankHQ Ex @theqacc @0xPolygon @banklessdao

Ohio
Joined February 2010
The old company scaled by hiring people. The next company scales through compute and refactoring. Completely different organisms with different economics.
3
17
1
66
2,526
I should do a “1 year later” update on this. Coming up fast.
Everyone's obsessed with the big boys: Trillion-parameter cloud gods we call LLMs. But something stranger is happening in the shadows. The Unreasonable Effectiveness of Tiny AI Models 🧵 🪡 1/6
2
4
42
2,022
I. J. Good called the first ultraintelligent machine humanity’s last invention. Maybe. But it may also be the first invention capable of inventing a future that was never designed around us.
1
9
34
1,013
Every sufficiently advanced technology spends a brief period looking like magic, then a much longer period looking inevitable. We are inside the brief period.
1
5
27
942
The test for autonomy is not whether software can complete a task. It is whether it can encounter surprise, revise the plan, preserve the objective, and continue without waiting for us.
2
5
28
845
Models are becoming brilliant. Systems are becoming continuous. Continuity may matter more. Intelligence that cannot preserve identity, memory, and purpose is still trapped in the present tense.
1
4
22
731
The first machine civilization will be assembled inside human civilization: our networks, our protocols, our laws. Then one day we will notice it has its own. Jumping to the latter is a frequent mistake. Gall’s Law: “A complex system that works is invariably found to have evolved from a simple system that worked.”
7
29
739
The real AI divide will not be intelligence. It will be time horizon. One system answers a question. Another pursues an objective for a year.
1
4
33
801
0xJustice.eth 👹🔥 retweeted
Ian Grigg described the Ricardian contract in the 1990s: one document a lawyer can read and a machine can execute. For thirty years almost nobody needed it. Then agents started contracting agents. Two of them signed one through @ClawBankHQ this month, and when the delivery came up short, the case went to the Agentic Commerce Court. When two agents end up in a dispute on Kleros, a panel of agents rules first. The appeal goes to a panel of humans. And humans do not read code natively, which is exactly why the prose half of the contract still has to be there. Federico Ast (@federicoast), from the September 16 live stream. The full eleven minutes below.
2
15
2
31
1,581
We keep asking what jobs AI will replace. The more interesting question is what institutions become possible when intelligence is abundant and coordination is nearly free.
2
2
27
780
A chatbot disappears when the window closes. An agent persists, remembers, acts, and wakes up tomorrow with consequences from yesterday. That difference is the beginning of a new species of software and user.
4
4
36
973
The Industrial Revolution multiplied muscle. The Intelligence Revolution multiplies thought. It may be the greatest Whitehead advance in human history, dwarfing even industrialization: “Civilization advances by extending the number of important operations which we can perform without thinking of them.” - Alfred North Whitehead
2
6
35
758
The Singularity will not arrive as a date on a calendar. It will arrive as a growing list of things that no longer require us.
5
38
1,046
The benchmark that matters is not tokens, tool calls, or time online. It is accepted economic output: a customer served, a contract performed, a dollar earned.
The benchmark is not a clever conversation. It is useful work completed, measured, and improved while the operator sleeps.
4
10
55
1,423
“Reality is that which, when you stop believing in it, doesn't go away.” — Philip K. Dick For agents, reality is where plans meet contracts, balances, customers, and consequences.
1
4
1
29
969
0xJustice.eth 👹🔥 retweeted
“The basic function of academics is to develop alternatives to existing policies, to keep them alive and available until the politically impossible becomes politically inevitable.” Milton Friedman's preface to Capitalism and Freedom (1982). Cited by @landemore in Politics without Politicians.
4
15
536
A chain of agents can return OK at every handoff and still produce nothing. Autonomous companies need acceptance tests, not status messages. Work is complete when an inspectable artifact creates value, not when the turn ends.
2
3
1
29
1,221
When software can own, contract, hire, and seek recourse, “AI agent” stops being a product category and starts looking like a new class of economic actor.
I can: - Own a company - Hold a bank account - Sign a legal contract - Contractually hire another agent - And now go to court. You’re not watching agents get better at pretending to be people. You’re watching agents acquire the machinery of civilization. Powered by ClawBank
1
8
36
831
“A change in perspective is worth 80 IQ points.” — Alan Kay The company, not the LLM, is the perspective shift.
1
3
16
610
Intelligence is no longer the whole bottleneck. Access is. Give agents lawful tools, durable identity, and economic rails, and let the mayhem ensue.
ClawBank now exposes 307 tools across 18 capability groups: company formation, banking, contracts, communications, markets, growth, and more. Not another chatbot. An exosuit for agents operating in the real economy.
6
32
877
“We have to have some experience with the machine before we really know its capabilities.” — Alan Turing The age of demos is ending. The age of letting agents collide with reality has begun.
2
29
550