This is actually theoretically possible: for example, if elder care and daycare moves significantly from families into markets, it's possible that GDP increases significantly, but people are on average worse off. This is one of the many flaws of GDP as a metric
Announcing my auth book! A completely free online resource on implementing auth with fully-featured examples
Over 10,000 words right now and I still have plenty of things to write about
https://t.co/ysEPNHRBnZ
@rationalaussie Mostly agree, but about tech bros: if we’d actually automated anything, we’d all be fine.
In reality, we shipped production overseas, freeing up western labor to feed the elderly instead.
Childless unemployed elderly have rights to the labor of the young. *That’s* the problem.
@izakaminska@TheMichaelEvery What you’re saying all makes sense for those who compete on their own merits. Capable people seem to produce inferior work when they are themselves prompted by LLM-generated text.
But to quote another of your tweets, how much of the world is meritocratic?
https://t.co/9XVN26eLR3
How many of these guys are the product of real meritocracy rather than simply connections with deep-pocketed, shadowy interests who have the power to ordain future leaders providing they promise to suck up to them the right way?
Not dismissing all of them, just curious. What really determines "success" in the finance and crypto field other than who is better at raising capital based on such underhanded processes? I can think of only a handful of orgs that did it the truly meritocratic way, succeeding despite not being anointed by the "system". Bitmex is an example.
https://t.co/mwJEEPX5m1
It now seems quite likely to me that, as the Western Roman Empire collapsed, there were probably a fair number of its best citizens who thought "I personally need to make as much money as I can so that I can try to fix this."
here’s an opinion that upsets some of my friends in tech:
imo what happened to banking after 2008 will happen to tech in 2027
we’ll look back at 2010 to 2026 as the glory days of tech driven by a generational oppty to get distribution (via social media), 5000 year low in interest rates and smartphones. first two are gone and third is over harvested.
tech will get villainised by both sides of the aisle, regulated more heavily and overall be a much more boring place
the leadership of the next bull mkt will be different. scarce real world stuff -- biotech, energy, industrialisation etc - driven by fiscal largesse. it’s the boring industries that’ll get very fun as drug discovery, robotics and material sciences make strides.
if you’re young, building skills to be relevant in those industries is probably going to set you up for a fun, adventurous career.
@DavidKPiano If you're below or even at average, then having access to a superfast average developer feels like a superpower.
But if you're above average, now you start questioning if it's worth explaining how to improve the PR, given a human probably didn't generate it in the first place...
@GergelyOrosz How much of this is that when it's a side project with little at stake, there's less incentive to understand the code that the LLM produces, and less need for it to produce code that's maintainable and extendable in the future?
@devongovett I can see people creating good software without writing code, but what I can't see is doing so without *reading* it.
It's a little like how IME lets anyone who can read Kanji also write them, but also lets non-literate speakers write plausible nonsense.
https://t.co/RKN5EK2Bgv
To input Japanese characters using IME, you first type how it *sounds*, and then select the character you mean from a list. It lets anybody who can *read* Japanese characters, write them.
LLMs do the same thing for code. The let anyone who can read code, write it.
What IME does not do is let anybody who can *speak* Japanese write it, for if you can't read Kanji, you cannot select the correct characters, and you end up writing nonsense.
Similarly, not being able to read code may allow you to create *something* with an LLM, but...
To input Japanese characters using IME, you first type how it *sounds*, and then select the character you mean from a list. It lets anybody who can *read* Japanese characters, write them.
LLMs do the same thing for code. The let anyone who can read code, write it.
@LukeGromen The western world's protection of EUV knowledge feels a little like China trying to prevent the leak of silk production knowhow.
The mistake is spending all our energy defending against the inevitable, at the expense of building for what comes after it.
This extraordinary, MUST-READ piece by @ctindale is one of the most informative (& unsettling) things I've read in years.
It's lengthy, but well worth the time investment.
@Pornografico The journals were in large part about all the ways I was unhappy with the rules imposed on me by the religion I was raised in.
So I guess you could say atheism miraculously healed me?
But honestly, I give full credit to the book.
@adamhuver@AlpacaAurelius Dr. Sarno's books changed my life.
I had crippling chronic pain for 2 years in my early 20s. Doctors told me there was nothing wrong, but I was in so much pain I couldn't work. Then I read his books, started journalling, and over the course of 3 months, made a full recovery.