Lesaun Harvey

@daniel_mac8 @sama Hey Dan, I’d say otherwise (with lots of nuance and thought) and love work.

@sama 100% AI will create a productivity and job boom. The people saying otherwise I suspect secretly hate working. Which, I get, but that's no way to live.

"post-AGI, no one is going to work and the economy is going to collapse" "i am switching to polyphasic sleep because GPT-5.5 in codex is so good that i can't afford to be sleeping for such long stretches and miss out on working"

Property does not begin with things already divided into claims. The world is given first: land, water, timber, minerals, energy, orbit, distance. Property begins when some part of that world is ordered into use; it becomes stable at scale when that ordering is made recognizable as a claim. It is not mere possession, nor mere physical control. It is control made legible within a system of recognition, exclusion, transfer, liability, and authority. In Locke’s account, labor belongs in that register. Labor is not bare exertion. It is the property-founding act by which human agency removes something from the common state of nature and orders it into use. Sometimes that act is performed directly by the claimant’s own hands. Sometimes it is carried through workers, animals, tools, capital, or institutional command. What matters is not the isolated motion, but the attributable ordering by which the thing becomes useful, held, improved, transferred, or preserved. AI and robotics bring this distinction forward. Machines can dig, lift, design, monitor, assemble, bargain, repair, route, optimize, and repeat. They can perform not only brute operations, but many sub-acts of judgment and coordination. As that happens, the important question shifts from who can perform the operation to whose ordering is recognized as valid. A road is not validly built just because machines can build it. It requires title, right of way, public authority, funding, liability, maintenance, and use. A satellite constellation is the same problem at higher altitude: automated launch and coordination do not substitute for license, liability, orbital priority, and accountable control. Property supplies the form that tells strangers who may enter, who must forbear, who may exclude, who may transfer, who bears loss, and who may authorize. Automation therefore does not make property less important. It makes property more directly operative. When execution is cheap, the scarce act is often the valid command. The economic consequence follows. AI may acquire overwhelming advantage in ordinary production, but advantage in one branch does not imply capacity in every branch. A machine may administer, manage, transact, and optimize; those are operations. They can carry out the ordering of things, but not supply the standing that makes that ordering finally belong to someone. The relevant human advantage is therefore not superiority in execution. It is standing within the order itself. Human persons are the beings for whom claims are held, transferred, inherited, and received. Because that advantage is institutional, capability alone does not defeat it. It is threatened only if the order changes whom it recognizes as a claimant. Human orders already create new claimants through posterity: children and descendants enter as future bearers of claim, and existing persons accept the cost of those claims because posterity is not external to the order; it is one of the reasons the order exists. Recognizing AI as persons would not continue that rule. It would authorize a second process of claimant creation, allowing instruments made within the order to become peer claimants against the inheritance they were built to serve. For persons who care about themselves and their posterity, that recognition dilutes the very inheritance the instruments were meant to enlarge. The question, then, is not only whether humans will still have jobs. It is whether automated production remains ordered to persons and their posterity, or whether the instruments built to serve that order become claimants against it.

@citrini It's inevitable, but the AI needs to cook.

Google should go all in on robotics imo

@liron @BrandonGoldman I wouldn’t bet against the economy bifurcating into real productive jobs vs. rentier/grift ones. I can imagine near-total white-collar disruption in 2 years—but mostly income-grifting adaptation instead of mass unemployment. Still, +2% rise seems plausible.

I claim (60% confidence) that 2 years will be enough time for data to show an early trend of AI pushing humans permanently out of jobs! So I’m happy to bet with Will at 1:1 odds. Yes, I already lost a similar bet in the 2023-25 timeframe against @BrandonGoldman, but I persist.

The point being: once an order is organized around persons, some of its central economic tasks are setting ends, authorizing technology, holding claims, and stewarding goods. Those tasks attach to beings for whom the order exists, not merely to whatever instrument is most efficient. In that setting, creating new persons within the order is an economic event, since it introduces new claimants against a structure already oriented toward existing persons and their posterity. So there is an economic reason to make better instruments rather than new peers, and thereby preserve a human comparative advantage in those tasks, since for non-persons the opportunity cost is effectively infinite.

@AbHomineDeus It seems likely there will still be comparative advantage in being able to do the tasks downstream of recognized dignity--empowering us to leverage technology toward our own ends, in relation with nature, without arbitrary interference.

People just really really don't grok what's coming. AI and robotics will do EVERYTHING we do better/faster/cheaper/safer. There will be no margin of comparative advantage. You will be economically irrelevant as anything other than a consumer. Make peace with it.

@krismicinski @moyix Oh, I did have an additional inference relating to the productivity increases more bug finding represents which is what the Senator questioned as hype.

@LesaunH @moyix 🤔 I disagree but don’t feel strongly about it

Excuse me I was assured it was all hype

@krismicinski @moyix What? There’s no reason to read Brendan’s “it” as not referring to the larger “AI is all hype” refrain just because he showed a concrete win in bug finding.

@LesaunH @moyix that isn't really what Brendan is saying, to be fair. His claim seems more specific to bug finding. No doubt that there are people who think that AI is generally hype--that's just not the specific item under discussion here.

@krismicinski @moyix x.com/atrupar/status…

@moyix was there anyone who actually said it was all hype?

I think many people would rather pay for access to something like country clubs as the ground of their relations than buy and sell their relations.

Dockworkers don’t seem to parallel bank tellers. So instead they demand corporations create inefficient rentier jobs.

@MatthewJBar Being early with the frontier still requires massive compute spend — letting the labs demonstrate it first and gather valuable capital before open-source catches up. Plus there are plausibly classes of models where open-source will stay years further behind.

@MatthewJBar I think they all hope to compute their way to being the first to sell a Kardashev 2 civilization and negligible senescence. x.com/sama/status/19…

What's the plan for big AI labs to become profitable when open-source AI is consistently under a year behind? They aren't profitable now, and they're under constant pressure to cut prices. Is the hope really that they'll initiate a software singularity and take over the world?

@Noahpinion The opposite. People need to reckon now with contributing to a shrinking slice of the economy and no longer having a critical hand in producing most of the goods their welfare depends on.

This. Exactly this.

Self-replicating systems suffusing the world with powerful ambient intelligent actuators responsive to local command is the archetypal end of techno-capitalism under consumer welfare.