I think it absolutely makes sense. ChatGPT's strength is how generalised it is as a tool, but openAI will never able to adapt the platform to every single use case. You can absolutely use it to learn a language for instance, but a great AI language learning platform needs a better tailored UI, it needs all kinds of non-AI functionality around it like idk a spaced repetition system, it might need to integrate into other platforms, and good prompting to be effective. AI isn't the product itself, but a component to try to solve a problem. And honestly I wish more startups focused less on simply "AI" and more on the problems it should solve.
If for nothing else openAI won't be able to market itself for every single use case, and so long as people aren't using chatGPT for some use case (even if it could perform the task) there's still an opening.
But any use case that gets large enough and makes money will be absorbed by openai direct, based on the market developed by the startup. OpenAI is using the Amazon model. Let someone else spend the money figuring out which market segments are profitable, then steal them with their inherently better access to the platform.
There are use cases that don’t align with OpenAI/Anthropic’s business model, which is to always use more tokens to get better. As the models improve they become way more expensive, with order of magnitude increases in price. I think there is a lane for detecting what can and should be deterministic after doing it a few times, and making it concrete so you don’t just burn tokens every time.
Their seminal book literally opens with explaining how most colonial institutions were created to extract the wealth of natives and slaves, institutions that live on today through the authoritarian regimes that inherited them.
Krugman is controversial at times, but he is a serious economist. The quote is also taken out of two important contexts. 1. It was not a serious academic prediction but rather part of a fluff piece by the Times about future precitions. 2. The quote also exists in the context of a debate that was raging during the advent of the internet whether or not the internet would hyper charge productivity and economic growth. In hindsight, especially after the late 1990s, the truth is far closer to the fax machine rather than a new era of prosperity. The lack of visible productivity growth from a technology that has so visibly transformed our society is one of the bigger questions in economics today.
The lack of visible productivity growth. In particular, it's not as visible as you expect in the GDP statistics.
Take TV Guide, for instance. Now, if anybody watches broadcast television, they can use the internet to find out what's on when. Is that better or worse for users than having a paper TV guide? In many ways it's better. But it shows up in the GDP as a negative, because nobody's buying TV Guide any more.
Or take Google. I can search for any information I want, for free. That creates immense value - immense in every sense except the GDP, where it doesn't show up at all, because it's free.
Wikipedia. Linux. gcc. The Wayback Machine. Even HN. All this is available to us, whenever we want it, for whatever purpose we want, for free. There's great value to us. Just nothing that shows up on the GDP statistics, because it's all free. (Yeah, I know, RedHat sells Linux, and Wikipedia asks for donations. They aren't Microsoft selling Windows and The World Book, though. You can still use them for free, and not get sued or jailed.)
These free things (Wikipedia, Linux, etc.) are sustained by a high GDP that translates to high incomes and, in turn, people willing to donate time and money for them to remain free.
Let's try living in a country with a measly GDP per capita of $1,000 (Rwanda for example) and see how many free things we can get..
People give more consistently in more pro-social societies. But those societies are poorer, with lower percentage of computer ownership and programming skills. So they give things like care and food.
Trying to spin the west as the place where altruism is most prevalent seems incorrect. More data needed.
I think we're in agreement here. I never said the West inherently has more altruism. However, the West's high GDP is what enables Westerners to give more to free computing/internet-related projects offering excellent value.
Without a high GDP, there won't be much to give in the first place.
Sure, there's a lot of value. There's also a lot of slops and negative value. These days I don't even use wikipedia that much even though I googled things constantly.
GDP is a very gross measure of things to be sure, but also difficult to fake.
Kind of like saying Lysenko was controversial at times.
He's not "controversial," he's just a bloviator. His credentials as a serious economist expired long ago when he signed up with the leftist team and agreed to never challenge them again, on anything.
When you have to spend many words on explaining the "context" in which someone's quotes should be viewed, you are losing.
> His credentials as a serious economist expired long ago when he signed up with the leftist team
You might be (unpleasantly) surprised by the emphasis of the current year's nobel economics prize winners on the importance of societal institutions and the need for inclusivity to advance the wealth of nations :-)
Krugman triggered some self reflection on my part when I read a column by him that was really stupid. Not just politically biased, but got economics wrong.
This is a Nobel winner in economics, I warned myself. Are you really putting your judgement ahead of theirs in their own field of study? I wanted to make sure I wasn't just doing the HN thing of arrogantly assuming I'm an expert in a field because I read a book once.
The eventual conclusion I came to was that I should always make sure to hold some allowance that the export may be correct and myself wrong, except when it's really obviously dumb, and that Krugman's writing fell into that camp.
> Kind of like saying Lysenko was controversial at times.
This is BS. Krugman has very specific theories that made predictions validated by practice.
Remember 2008? He made a prediction that an increase in the monetary base wouldn't cause inflation. This prediction was spectacularly confirmed. He also made a case for fiscal intervention: it wouldn't cause inflation, and it would speed up the recovery. And his prediction again was confirmed.
More recently: he predicted that the inflation spike was transient, due to supply chain issues rather than fundamental changes. And he's again been vindicated.
Saying that Krugman was vindicated is quite a stretch. As the economist Noah Smith wrote:
>...In 2021, Krugman tweeted: "I like it and plan to steal it. This report does look like what you'd expect if recent inflation was about transitory disruptions, not stagflation redux".
As Smith pointed out:
>...But in late 2021, inflation spread to become very broad-based. Services inflation was always significant, and took over from goods inflation as the main contributor in 2022.
>The notion that this was just some transient supply-chain disruptions that was only affecting specific products was absolutely central to Team Transitory’s claims in the summer of 2021. And that was incorrect.
>...Team Transitory also called the end of the inflation at least a year and a half too soon.
On October 13, 2021 Krugman tweeted "Three month core inflation. Why isn't everyone calling this a victory for team transitory"
>...So they didn’t entirely whiff here. They just greatly overstated their case. And their complacency in 2021 probably fed into the Fed’s decision to delay the start of rate hikes until 2022, which in retrospect looks like a serious mistake.
What did get vindicated was mainstream economics as taught in our textbooks. As Smith wrote:
>...Mainstream macro’s first victory was in predicting that the inflation would happen in the first place. In February 2021, Olivier Blanchard used a very simple “output gap” model to predict that Biden’s Covid relief bill would raise demand by enough to show up in the inflation numbers. His prediction came true. He didn’t get everything right — he thought wages would rise more than consumer prices, and he neglected the lagged effects of Trump’s Covid relief packages and Fed lending programs. But his standard simple mainstream model got the basic prediction right when most people made the opposite prediction, and this deserves recognition.
>More importantly, mainstream macro appears to have gotten policy right.
Do you actually think China has fairer courts, more open financial markets, and better business environment than the EU or US? For the longest time Hong Kong had to fill those gaps for China in order for the country to attract capital.
Regardless, even if China does today the explanation still holds considering China only transformed a few decades before the book was written.
I think China is a worse country to live that the US or EU, but a better business environment. China's State Capitalism is simply more efficient than the neoliberalism hellscape that came up from the Soviet Union's collapse.
I see the rest of the world becoming more similar to China than the other way around, unfortunately. The US is already a pseudo-Soviet state economically, and is on the brink of becoming a dictatorship on top of that.
Extremely inequality that keeps increasing, political movements that try to change this are killed in the crib, a dying empire.
Just because the Central Comittee is called Blackrock, doesn't mean the American economy is not being more and more centrally planned. Free Enterprise is dying.
I mean economically, the political system is still "free" (less and less), it's not a dictatorship (yet).
If for nothing else openAI won't be able to market itself for every single use case, and so long as people aren't using chatGPT for some use case (even if it could perform the task) there's still an opening.