Because the profits from any company which _could_ trivially be redone by an OpenAI agent will crash to nothing, because it could trivially be done by someone else's agent too. It's what ultimately makes the economics of OpenAI, Anthropic et al so crazy: they need to capture trillions of dollars in value, but those trillions of dollars won't exist any more. What's the value of writing a piece of marketing copy now, the same as it was last year? That, across the whole economy.
One might say, in the end, what's wrong with being self-serving? Are we _all_ meant to be self-sacrificing? Are we all meant to be the eggs that get cracked to build this wonderful future ommlette?
Has OpenAI shared any new interesting _problems_? I.e. do more than strip-mine mathematics, as said in the article, but said: "Here is a question I cannot answer, it's interesting for this and that reason, I believe it's true for this and that reason, it's beyond me though, have at it." Has that happened yet?
Fair, in that no analogy is going to be very good given that this really is something new in the world.
Consider anti-money-laundering rules. BoA just paid more than a billion dollars, not for money laundering, not for telling others to launder money, but because they should have tried harder to stop others from laundering money. Intent is not everything, knowing that your system _could_ be abused and not doing enough about it is enough, and that seems to fit OpenAI rather well.
If I leave my computer open to the internet and some other group uses it for nefarious purposes, yes generally I would not expect to be held at fault.
If I *train* that other group, if I teach them how to hack computers and give them incentives to do so, and then I leave my computer open so they can use it for whatever purposes they chose, then I would expect to be held at fault. The maker is responsible for that they make.
This sounds irrelevant but it isn't: I just tried to unsubscribe to marketing emails from some Meta property, and got:
Please confirm that you want to unsubscribe.
Unsupported request - method type: post
That's how careful this company is. Who knows how long that flow has been broken for, no one in the company is motivated to care, because what impact can that possibly give? How could that look good on your PSC? No company has a worse history with user privacy and no company is less deserving of the sort of trust using Muse requires.
I, and I'm pretty sure I share this impression with many of you, would feel a whole lot more optimistic about AI for software engineering (which truly has been transformational) if it didn't come with layoffs or the promise of layoffs.
Subscriptions do seem to continue to go _up_, too. The proof of the productivity pudding is lower prices, which means the proof of all this AI hype is not so much Millennium Prizes being solved than subscriptions actually getting cheaper. If AI is already having a huge effect, inflation should be going down... if AI will have the effects OpenAI et al are currently selling, prices should plumet. I mean, I'm waiting...
At one point reading this I thought of Molly's Soliloquy: "...frseeeeeeeefronnnng train somewhere whistling the strength those engines have in them like big giants and the water rolling all over and out of them all sides like the end of Loves old sweeeetsonnnng the poor men that have to be out all the night from their wives and families in those roasting engines stifling it was today Im glad I burned the half of those old Freemans and Photo Bits leaving things like that lying about hes getting very careless...". You probably don't want to recall Joyce if you want to get people to understand what you're doing.
Fun story: at my previous aaaawful company CC was discovered as a thing to care about at about the same time as PMs and managers were encouraged to land code changes using the _then_ quite terrible AI tooling (this was a year or two ago). Cue an avalanche of completely unreviewable diffs.
I think this is a counter (a cynical, scary counter to be sure) to the charge that OpenAI/Anthropic etc are unsustainable and will run out of money and it's all going to fall down. These companies have amazing, unprecedented power.
I can't say it doesn't make sense. Who still loves Facebook? Older people. Who loves chatbots? Old people too. Not _just_ old people of course, but this genuinely might be the first new tech I've seen which has been so easily and quickly welcomed by those a generation or two older than me.
Why hasn't AI already solved the chip shortage? I guess we need to spend more on AI. More money to the AI barons so they can solve cancer and also employment.
Lots of "performance review boilerplate, I don't want to say anything _bad_ about this person but there's nothing particularly good I can say" energy there.
Someone decided that this feedback was worth highlighting:
"Bob demonstrated a level of intelligence and contextual understanding that goes beyond anything I've seen in other tools. Bob goes beyond anything I've seen in other tools."
Skipping past the implication that the author believes that the only things that matter are the things he personally cares about, he doesn't seem to realise that we use the word "subjective" to mean two different things. We use it to mean _not objective_, and we use it to mean _internal_. In "subjective experience" we use the internal meaning of the word. Subjective experience is a fact, it is indeed the only fact that we each can know for 100% certain, all other facts are inference and induction and guesswork, but you cannot deny that you have subjective experience. It is _objectively true_ that I have subjective experience. The author wants to contend that it is _subjectively_ true that I have subjective experience.
To take an example, I don't know whether my cat has subjective experience, and I think it is impossible to know for certain. Nevertheless, there is a fact of the matter, it is either true or false, it is not a matter of my care for the animal, it is not in the eye of the beholder, it is an innate property that a system might or might not have.
And civilizations do fall. The bronze age civilizations did collapse. There was something dark about the dark ages. There is no law of nature which says that things will be okay, and plenty of examples where things have fallen apart.
Just to add to this, if AI is as epochal a change as its boosters say (and I am open to that argument), then surely it must have the potential for awful consequences, in proportion to its potential for good.
This is kind of damning, isn't it? "Similar or slightly higher" for spending the _same time_ means it's not, in fact, helping. At the same time as being extremely damaging to a significant number of students who, of course, use the AI to cheat. Because everything about it makes cheating easy.
I do wonder about intent. Meta has spent 20 years paying its engineers and others vast amounts of money to Increase Engagement, and punishes those who fail to find ways to do so (over the past few years, by firing them). And it's worked! Engagement has increased greatly over the years, Meta survived the effective end of Old social media (you know, the stuff from your friends) and thrives in the new world of infinite content. Was the intent to _addict_ people? It doesn't have to be, to create a system people have become addicted to. Even if the harm is not intended, does that mean it cannot be regulated? Surely not! Let's ban infinite scroll for a start. The question of whether to punish Meta is quite separate from the question of how to reduce the harms of its products.
Either people are being really, really silly (which cannot be discounted), or the potential reward is so high as to override whatever qualms a normal person must have. Is that it? Is this people looking at a solid career at Apple or sudden millions from OpenAI, and thinking the risk is worth it somehow? Or, more darkly, is it people thinking _this is my only chance and I have to take it_? Or is it trickle-down lawlessness?