Because the university departments are held by a certain type of individuals who want to play down and dismiss anything held sacred and holy by tribes, old cultures, religions, etc., explaining them away with "totems", "magic rituals", etc.
This article is ultimately about support for Jordanes' claim that the Goths originated in Scandinavia and then some new findings that the Goths spread by gene flow, not just cultural assimilation. Jordanes was a Christian writing among a Gothic community that had converted to Christianity, and they had no attachment to whatever was "sacred and holy" to their ancestors centuries before. Consequently, historians expressing scepticism about Jordanes claim were not doing it out of any bone with "sacredness and holiness".
Perhaps before trying to depict an ultimately rather minor controversy among historians as some axe ground against the sacred and the holy, it might be worth actually learning about the topic first?
In all honesty almost all exams are gamed by studying question patterns, studying previous years' papers, etc. There's nothing wrong in that whatsoever.
BUT it is such a toned down, dwarfed down version of what a holistic education should look like. It probably looks like that in undergrad CalTech Physics, maybe. Not much in most places.
People outside think it's so high and holy to be a PhD. It's not for most people.
Most people also go to academia because they can't find other jobs.
Things in Humanities are much worse than STEM, as there are much more subjectivity, favouritism, nepotism, groupism involved.
'Research' nowadays in most fields and institutions is nothing like whatever grand vision people tend to have: Faraday or Bell Labs, Maxwell or Bohr, Feynman or Bletchly Park.
You might be assuming that the skill is standalone and that the market will see a goldrush that drives the price to near zero.
An analyst could learn enough SQL in a few weeks to make their workflow better (leaving aside that AI analyst is already a thing).
Almost anyone could spend a few weeks getting a heavy truck license. Barriers to entering this trade are not that high but earnings have stabilised above subsistence level.
I'm sure there are plenty of analysts already using chatbots to improve their workflow. And the barrier to entry to heavy trucking isn't the license, but being being willing to not only drive a massive friggin truck, but to do so for many many hours a day often entailing being away from your family/relations for the vast majority of time. It's a tough life, and that's the reason the stereotype of truckers is what it is.
By contrast the LLM 'skill' of 'sit on ass and talk' is something basically everybody already does. So you just need some minimum level of general cognitive ability and you're good to go.
I appreciate what the project is trying to do, but why go on radical hyperbole such as "That, Ma says, is the essence of AI colonialism — the extraction of difficult, low-paid labor from the Global South for profit in the West."?
Also, do Chinese companies pay a higher rate? I have no idea. But Chinese data annotators also have to work with graphic content, I believe.
The workers are made to read bad things as part of AI training process, but they are adults agreeing to work in those things in exchange of pay, and they can always quit. The "workers" are computer literate English-speaking elites with stable internet connections even in poorer countries.
I have interacted with students from all over the world on different occasions. If I meet a student from India/US/China/Europe in a global research/academic/engineering forum, there is a non-trivial possibility that they are from a middle class or lower background, but when I meet, say, Africans, Bangladeshi people in the same places, they are almost always elites in their respective countries. It's not surprising, tbh.
My point it, when you are working for 4-5 USD per hour from Africa, you can communicate using good English, have a decent computer, have stable internet connection, a college degree, you are doing the AI work by choice.
Being paid by hours even in India is a great privilege where the culture is to make people work 10-12 hours shift for the pay of 8.
They say, "$2", I don't have any direct experience, but I saw ads for $4-$5 per hour, and that is for non-specialist work, and if you get steady work, you are paid more than "engineers" with 4 year degree at the end of the month.
In India, freshers Engineers in consultation shops get ₹30-40k per month, and have to be in office for 8-12 hours per day. Where 4$ per hour pay will get you 4$ * ₹95 * 8 hours per day * 5 days per week * 4 weeks per month = ₹60,800, which is enough for a comfortable middle class life (outside metros like Mumbai, Bangalore, etc.). There is also no money to be paid for rent, commute, etc.
People getting into these positions are getting in by choice, and remaining by choice.
If they wanted to shed light on real woes, they should have focused on the small terms the workers are hired for. And work not being regular, people laid off on short notice, short term contracts, etc.
Seeing salary as well as prices in US dollars eyes is a common folly. This is what I believe is called "geo arb".
I think, instead of 'Birdwatching', it could be any activity that required uninterrupted long time engagements, no immediate rewards, active involvement, slow progress, etc. Like cooking, painting, making films, writing, etc. You get the idea.
I have tinkered with Pharo before, but I have come across quite a few entries on HN front page where I encountered projects, releases and I had no idea what the thing was! It could be anything from a networking library to LLM wrapper, from a music band to an programming language.
Clicking on this page, one finds:
> What is Pharo?
>
> Pharo is a pure object-oriented programming language and a powerful environment focused on simplicity and immediate feedback.
I really appreciate these one liner introductions on any project.
I wish this was somehow listed as a separate line item with the price, assuming prices will normalize at some point. I don’t want these prices to be considered the new normal, even after memory prices come down. I think transparency in the pricing would help with that.
I developed many models from scratch deployed in various fields.
When you want to automatically detect a pathology from a give scan, you want it to be model that you own, train, use, and update, and not a generic vision capable LLM.
This is NOT what benchmaxxing is. You want the model to be accurate in terms of your cared metric- depending on the problem.
reply