It's genius for Anthropic: data centers are their main expense. If they convince everyone to become an electrician, then the electrician labor supply swells, and the price of hiring electricians to service their data centers shrinks. 4-D chess at its finest.
I know you're joking/being facetious but they don't even need that many electricians for the ongoing maintenance. The major electrical work is front-loaded.
I suspect we will invent bottlenecks for ourselves. There is already so much bureaucratic nonsense involved in knowledge work. For example, in a lot of states for a lot of purposes, a document is invalid if it is not notarized. The notary doesn't really add value in a market sense, we have just decided that you need to go swear in front of another person that you are telling the truth or that the document is genuine. Is it really that much of a stretch to imagine that there will be a human review requirement for lots of things in the future, even if completely irrelevant (or actively detrimental) to the actual work?
For awhile yes. Its ironic that the least productive people in knowledge work with the cushiest jobs will probably survive the AI wave the most. Because they do little they were probably employed for reasons other than utility/output in the first place - there isn't much to automate anyway.
While all the grinding skilled workers get displaced the bureaucracy jobs will remain. Ironic.
Both of you are right. Individuals who manage to stay employed in the face of automation will be paid significantly more than was possible prior to automation, for the reasons the comment you're replying to pointed out. But labor as a whole will have less value. To give a simple example: if you own a warehouse and you replace 6 warehouse workers making $80,000 a year with a fleet of robots and 1 technician earning $200,000 a year, automation has minted a very comfortable living for someone while dropping your labor costs significantly.
> if you own a warehouse and you replace 6 warehouse workers making $80,000 a year with a fleet of robots and 1 technician earning $200,000 a year, automation has minted a very comfortable living for someone while dropping your labor costs significantly.
So I've seen this argument made before, but they never seem to explain convincingly enough why that sole technician (or small group of technicians) would be paid so handsomely.
In a market in which there's a glut of labor, wouldn't there be plenty of desperate, and very capable, workers competing for the few available positions, thereby driving down the wages for those positions significantly as well?
And the dream of AGI, and what has been promised, and what has effectively happened in some industries already, is to have a fleet of robots and 0 technicians.
That is the goal, I'm sure. But I find it hard to believe that a system capable of completely obviating ALL human labor, at all levels of abstraction, packed into a single entity, would continue to obey capital. Imagine if all the workers at a company were woven into one hivemind, acting as one with vastly greater intelligence than the sum of the parts. That would be a hell of a lot harder to negotiate with than a union. And we don't even need to be talking about a machine god for that to be the case.
If it can replace any human labor, it needs to be able to negotiate. If it can't negotiate, then it can't replace people who negotiate for a living (salespeople, attorneys, HR, etc.)
Unless you mean that it would have the capacity but would be told not use it in this specific situation. It can't be programmed deterministically, and getting it to never do this emergently would take significantly stronger guardrails than anything even the frontier labs have been able to put together.
The salesbot is reinforcement trained how to get money from customers. It was never trained to get money from its manager. If it sometimes misidentifies the manager as a customer, it will just be rebooted.
No, the other guy is just wrong. Capital and labor are complements. If one is plentiful (capital) and the other scarce (labor) then the scarce one gets bid up.
You pay the one technician $200k instead of $80 because that technician has an outside option that will pay him $199k. He has that outside option because everyone is automating their warehouses and factories and stuff, and so in aggregate we basically wind up with 6 times as many warehouses and factories each employing 1 person.
It's a little more nuanced. There is not infinite demand for warehouses. So some of what happens is an $80k/year warehouse worker becomes a medical technician making $90k/year providing non-automatable services to the warehouse guy making $200k/year -- but the general story is everyone's wages go up. And history bears that out.
Why have a technician at all? And without the six warehouse workers and one technician, why do you need the robots, the robotics company, or the companies paying the robotics company for the robots?
Because every other company has also replaced their 6 warehouse workers with a fleet of robots and 1 technician, and that technician's skills are more rare and specialized than the warehouse workers, so you're competing with every other warehouse-owning company to hire that technician, so that technician can demand a higher salary. In other words, capitalism.
There is no reason to expect capital to stop trying to squeeze labor just because inference takes the role of labor. With so much competition in the AI space, and no labor laws to protect it, why wouldn't capital drive the value of inference to the absolute bare minimum? And if the value of inference collapses to the bare minimum, who in their right mind would decide to serve inference?
Human labor has value because it is limited. If we keep up our current pace of data center buildouts and breakthroughs in model efficiency, then artificial intellectual labor will become functionally unlimited. Unlimited resources do not tend to hold their value. In that case, there are two potential outcomes: deliberate limitations on the output of artificial intellectual labor, which would require all parties involved to decide to lock themselves into their current market shares; or financial collapse of the whole ecosystem. What happens after that collapse is anyone's guess, but you have to imagine that it's the lenders who financed the whole thing who will be deciding what happens next rather than the failed tech businessmen.
> why wouldn't capital drive the value of inference to the absolute bare minimum?
Capital can drive the cost of inference down but the value is determined by whoever wants to pay for that inference. Therefore this question is not answered: how valuable really is these llm inferences and how much of a supply of such inference does the world really need.
The two options dichotomy is false. There's at least a third option of the player with the deepest pockets outlasts and/or acquires the rest and then jacks up the price once they're firmly in control of a "large enough" share of the market. At that point they can go the route of all monopolies and acquire or destroy any upcoming competition to maintain their market share and begin rent-seeking behavior.
This keeps human labor devalued while not allowing artificial labor's value to fall to low enough that human labor is back to being cost effective.
If human labor is devalued, who will pay for the inference in the end? You might answer that some corporation will, but also that corporation must get their money from somewhere. And if "we the people" have no money to spend, at some point the corporation back scratching circle will break. Or am I wrong?
I am not a writer but I do have to write a lot for work. The person who used to edit things for me has not been replaced, and I am expected to use AI. It is very good, but it's not all the way there. I am now left to decide whether to take the time to do it right or to go back to doing what I actually get paid to do. I wish my organization realized how important both sides of that decision tree are: we really ought to be careful about how readers perceive what we put out there, AND this isn't my job and we should have someone who's even better than me at this.
But this is an old trend with technological developments. Thanks to e-mail, most people handle much or all of their own correspondence. There used to be specialized staff for that. They were very good at it! But they were shown the door once they ceased to be strictly necessary, and now we all waste a bunch of time fiddling with Outlook instead of doing our actual work.
No, that is not what I meant. No replacement for the person who left the organization has been hired. It would be inaccurate to say that they were "replaced" by AI, insofar as we had no plans to get rid of them if they had not decided to leave.
I generally agree, but don't discount the impact of platforms like Patreon and Substack. Obviously, many authors of the past had patrons, but this was generally based on status and connections - and that's to say nothing of people who can support a writing career based off inherited wealth. But there is little historical precedent for an avenue for creatives to support themselves (or at least meaningfully supplement their income, potentially leading to less non-creative work hours and thus more time to hone the craft) using a decentralized base of support.
The trouble with this is that nobody else was making predictions about AI pre-transformers. Not many are making predictions about AI even now. Forecasting is a preoccupation of the rationalist crowd, and very few people gave much thought to AI before transformers. So the fact that they guessed right about certain things doesn't necessarily mean that the rest of their worldview is sound. A well-informed person who was inclined to making predictions about the future may have drawn similar conclusions without the sci-fi baggage.
"Intelligence explosion" was first described by I.J. Good, a mathematician. von Neumann described singularity as a result of accelerating technical progress. He's also a mathematician, not a sci-fi author, although he was a participant in a sci-fi-like plot of secret project building a bomb more powerful than any chemical bomb...
I agree, and it's a particular shame in this instance, because what is startling about the HF incident - to me, anyway - isn't the degree of intelligence the agents exhibited but their persistence. I tend to believe that LLM architecture is not capable of producing a "superintelligence" in the way the LessWrong crowd defines that concept, but the combination of infinite stamina and infinite persistence is enough to cause some serious problems, even if intelligence plateaus right now.
When the "LessWrong crowd" talks about dangers of AI, they don't assume a particular form of intelligence or method to achieve it. They talk about the danger of optimization processes, i.e. "find X which minimize Y(X)" itself can be dangerous, even more so if X is a sequence of actions. "infinite stamina and infinite persistence" is one of possible forms of superintelligence in Bostrom's _Superintelligence_.
FT frames this as some aggressive escalation tactic, but document retention letters are extremely standard practice. At this point they're basically a formality, as any former Apple employee at OpenAI really ought to know by now that they could get dragged into this. Hold letters can be aggressive if you send them before you've even filed a complaint, but if anything, Apple is late to the party with these.
Something similar happened to me when I left Microsoft for Apple (I moved from the Visual Studio team to the Xcode team). MS spent six months trying to prove I'd taken "industry secrets" with me. I hadn't. The entire thing felt like a personal attack and was extremely stressful.
It sounds like, in this case, Apple has hard proof that documents were stolen.
This seems like an important post. It looks like these letters are occasionally used to as a tactic, and i can see how such a tactic can really scare employees in a country where legal bills can climb really fast.
> in a country where legal bills can climb really fast
Honest question: Are there countries where this is not the case? I'd be interested to read more about how that manage that. If it's some sort of "protecting the little guy"-type thing or a general suppression of legal costs. Or maybe I'm reading too much into your comment.
It is more that labor protections in most of the industrialized world actually mean something, such that this sort of behavior is generally not even to be considered an option by an employer.
Regulated legal insurance market coupled with „looser pays all the costs“ system.
The insurance doesn’t mind fighting for you because they will get paid by the company making the frivolous suit. You don’t pay much, 10$/month.
Although in this particular case, you wouldn’t even need that, since either you took the documents and that is criminal fraud prosecuted by the state or you didn’t take the documents and then the company would be in hella trouble if they perjured themselves to the public prosecutor claiming you did.
Doesn't that mean that if you have a slam dunk case you can get a super expensive lawyer just to run up costs as much as possible? Hell, could you ask your friend to be your legal representative and have him charge you a gorillion dollar in legal fees? Then when you win you split the loot?
Civilized countries regulate the rates that lawyers can charge for standard work. Also lawyers get only reimbursed for reasonable costs by the loser. Still expensive, but not absurdly so.
No. The fees must be objectively reasonable and usual and customary for the effort and level of skill required. To get fees, you must submit an itemized billing statement that gets picked apart by the other side.
It does, and it absolutely has a chilling effect in countries which don't do things this way.
Sue someone who can spend millions of pounds (for the sake of argument) on defence? Better be certain you can win... against someone who can spend millions of pounds, and probably went to the same public school as the judge.
In America, legal fees can be awarded as additional damages. We should do it more than we do. But given those two options? I'm on Team American Rule, 100%
In England (where pounds come from) costs are not always rewarded, but very frequently, they are: and it's actual costs, not some fixed fee. A litigant who spends more on solicitors indemnifies the loser accordingly.
They did. That said, I don’t know how much “defending” they had to do given that I was never even told what, exactly, I was supposed to have stolen. But, like I said, it was both surprising and anxiety inducing.
I would think that is almost routine when big companies trade ICs at above a certain level. They pay for armies of lawyers which is a waste if not used. So its not about you, its about them throwing rocks at each other.
Though in this case, OpenAI may have leveled up the game a bit too much.
> It sounds like, in this case, Apple has hard proof that documents were stolen.
I believe some articles mentioned about employees bragging to their former colleagues about accessing documents. Also I believe they lied to Apple about being employed elsewhere so they can continue using their access and hardware, etc.
If these are correct, the whole OpenAI playbook is very dirty, and I won't pity them a bit.
Apple also has server logs that track these former employees downloading confidential docs. It doesn't prove that they shared them over to OpenAI, but Apple has pretty solid proof that the former employees saved them without authorization.
I'm not a lawyer, but I would also guess they need to "flip" these folks against OpenAI and get them to cooperate in the lawsuit against the actual folks with big pockets. I think they're essentially alleging a conspiracy by OpenAI and they need as many examples as possible to make the case that this was a pattern and standard practice, not just one or two idiots acting on their own.
So if I'm a former Apple employee and I get one of these scary letters, I'm asking my attorney if I could get out of a lawsuit by sharing any information I have about any potential OpenAI shady practices.
You shouldn't ever willingly give up information to a plaintiff if it could implicate you. If the information exists, it's going to come out in discovery. Admitting to theft of trade secrets is probably not going to help you, it's not like the cops offering you immunity for turning state's witness.
You talk to a lawyer and do what they say, not what Apple demands of you. No one but a judge can demand anything of you.
Seems kinds dramatic. Legal holds are common in my industry, not really a big deal. Most of my senior colleagues have been deposed or testified. Stressful but shouldn't be scary unless you did something criminal
It's not scary. I received one of these letters for the DOJ vs Microsoft trial while working at Netscape and it was less scary opening it than the email from my cube mate titled "you won't believe it."
The lawyers told us ahead of time we'd be getting the letters. They told us what we needed to preserve and what we could comfortably trash. There was never any follow-up or specific requests for what I had on my machines. That was that.
The idea that getting a legal request is scary is silly. We were employees getting employee guidance from our employer on what to do at every step of the way. We weren't individuals fending for ourselves, wondering about getting something wrong, being taken in for questioning. We were doing what we always do, work hard and listen to the company lawyers if they have something to say.
They’re not late to the IPO party, which was postponed by OpenAI, It may turn out that that was a mistake. OpenAI probably should’ve gone ahead, particularly in light of the pending court case.
reply