Professional use of AI for programming is on the rise. Some employers are
forcing it on workes, while others are just making it available. I did spend
some time using it as I was transitioning from Perl to Golang and used it like
a lazy search engine to learn how to translate concepts from Perl the Golang.
The main value I saw was the Copilot Chat inside of VCS meant I didn’t need a
browser. There were fewer distractions available than if I had done the same
search in my browser.
I don’t think the claims made my proponents are real. I don’t think the claims
of organizations with billions of dollars invested in AI are accurate. My
experience thus far has been in line with other seasoned professionals. It
makes mistakes like an intern, only at a higher velocity. While AI may make
some tasks easier, I have serious concerns about its use professionally.
You can’t escape it. AI is everywhere. It’s proponents extoll god-like
properties and infinite capabilities upon it. There seem to be two camps, one
that believes the current generation of LLMs provide a pathway to AGI, and those like me who are growing more and more
convinced that GenAI is approaching the
bounds of what’s possible. It’s true that some of what these current LLM’s are
capable of is impressive. What remains to be proven is any sign of novelty or
comprehension.
We don’t yet have clear answers or strategies for the first and further order
effects of AI on our planet, economy, and society. The Accelerationists don’t
give a fuck. They believe if the system is going to break, it’s better to
break now than in 10 or 20 years. If AI could destroy humanity, they want it
to happen as fast as possible. This may seem like a crazy position, but they
believe there’s only two possible futures: one where AI solves all of
humanities problems and we live together in utopic bliss, or one where AI
becomes sentient, determines humans are useless and kills us off. Why wait to
find out which one? If were in for a utopia, we can’t get there fast enough.
If it’s human annhilation, well, it’s inevitible. If the human population is
increasing, the ethical choice is to speed up the apocalypse as to spare those
future masses from suffering.
Both of these outcomes require the current generative AI to lead to AGI, which
seems unlikely. More mundane, but no less problematic scenarios are more
likely. I have no faith that any technology will create an egaliterean
utopia. Nothing in the course of human history indicates we’re capable of
living happily together. Technology will not fix this, no matter how much
venture capital is available. The death grip capitalism has on tech to
squeeze every last cent of value from a consumer doesn’t select for utopian
ideals.