I just want to call out that the craft in the webpage is exactly the kind of thing AI cannot produce: it's quirky and tasteful and combines modern web influences with graphic print traditions - it is awesome.
I think we need to unionize, across companies. We need to be able to block stuff like this, and to be able to demand that you can’t lay off someone to replace them with AI (bringing US workers rights up to the bar set by China). We also need to be able to hold our leadership to some kind of code of ethics. I don’t want to work for a company that makes kill bots, or can renege on climate pledges.
The missing bit is a representation of knowledge, and a way to represent a learner’s comprehension.
Even if you shortcut by synthesizing a textbook in every major topic - that’s just one arbitrary representation, and the way topics overlap is outside of the material.
I am very interested in this though, if anyone has references to relevant research I’m all ears.
This is US centric, but you can show this to anyone in any major city in the western world and they will nod.
The problem is global. Any time you are told the problem is immigrants in your country or that your local politicians didn’t build enough housing - that’s not the root cause.
The root cause is the inevitable centralization of all wealth in corporations who can dodge taxes. This affords them the ability to accrue assets endlessly, buy politicians, achieve monopolies and bump up the prices.
There are 3 things that need to happen to restore the first frames:
1) some things like housing and healthcare must not be asset classes. It is unsustainable to demand profit growth year over year from either.
2) the wealth needs to be redistributed. It is a systemic risk for corporations and individuals to have accrued as much wealth and power as nations (and it doesn’t seem to be good for the mental health of those at the top)
3) we all need unions so workers can advocate for themselves and even the playing field. Including tech.
I’m not a communist, I think capitalism broadly works, but it requires regulation. We deregulated and it’s made a mess, we need to wind that back and patch the bad logic.
I didn’t mean to be hyperbolic, for carbon footprint I’m looking at cumulative footprint since 1750 and not only recent annual footprint. I’m glad we agree, I take your notes.
If you don't count the externalities, sure. Healthcare is a cost too. We need more holistic accounting, the financialising of everything into a tidy but ultimately false P&L column is literally killing us.
But does it work? I’ve used LLMs for log analysis and they have been prone to hallucinate reasons: depending on the logs the distance between cause and effects can be larger than context, usually we’re dealing with multiple failures at once for things to go badly wrong, and plenty of benign issues throw scary sounding errors.