TheOtherHobbes
IME LLMs often make reasonable assumptions, but sometimes they don't and you have to backtrack and say "No, what I meant was..."
When the process works it's almost magical because the LLM successfully anticipates your needs, and implements your request in a much fuller way than you might have done yourself.
When it doesn't it makes a mess you have to clean up.
But generally I've been impressed by how well the leading LLMs will Do The Thing from a brief description, interpolating useful and correct details without being told to.
Compilers don't work like that. They're translators, not (entry level) inventors with independent agency.
That was somewhat true when execution was time-consuming and expensive.
But original ideas are very far from worthless. They're very rare, because understanding domains and relationships at a deep level is a rare skill.
Meanwhile AI makes execution much cheaper and easier than it used to be.
The real reason is the models just aren't good enough.
Running a single successful fully automated business at scale would be a huge milestone and a legendary PR achievement, but the tech just isn't there yet.
We've never solved the fundamental problem, which is that there are two kinds of human. One is oriented towards peace, cooperation, and mutual aid, the other is sociopthaic/narcissistic and murderously selfish.
These two modes interact through history.
Traditionally the second kind has the biggest impact, because they're oriented towards violence and domination - good for short-term goals of personal indulgence, ambition, and expansion, but catastrophically destructive in the longer term.
On this cycle the second kind has literally started to destroy the entire planetary ecosystem. It's not obvious there's going to be another human cycle after this.
It's still an implementation of justice, and relies on a bedrock conception of justice as moral goal.
We don't agree with the implementation, but we still share the goal in the abstract.
Applies to STEM generally, because the foundation of STEM is peer review and independent testing and confirmation.
Even if the practice never quite lives up to the theory (true) that's no excuse to hand-wave away the requirement as if it doesn't matter.
Proofs that can't be confirmed are valueless because no one learns anything from them. The value of a proof isn't just in proving something but in inventing new techniques and insights that can be reused for other applications.
Brute force searches are fundamentally less valuable, because the creative foundation of math is discovering leaps of logic that solve problems with insightful shortcuts that can be generalised and reused for other applications.
We don't know the mechanism which generates the gauge symmetry that makes charge discrete or generates conservation laws from symmetries, so it's risky to be too prescriptive about what is and isn't possible.
In the limit you can imagine a meta-technology which manipulates the mathematical structures that generate reality instead of just assuming the rules are fixed. That might or might not be possible, but the rules have to come from somewhere and there's no absolute guarantee they're hardcoded.
This is already happening to crops like cocoa and coffee, and crops like maize and wheat are next.
If we're too stupid as a species to understand this, we can just let Darwin clean this up and give the planet another chance with different species in a few million years or so.
This is not hyperbole. If nothing changes that's a completely realistic outcome within 50 - 100 years.
See also:
https://www.theguardian.com/world/2026/aug/27/what-caused-th...
No matter what AO3 thinks, in many countries this immediately crosses the line into criminal harassment.
As for the rest - the fanfic community has been among the loudest criticising AI writing for 'stealing from real artists', which is... not exactly self-aware.
We used to have something resembling a democracy, now we have an evolving neo-fascist hellscape created by a cabal of the least interesting tech people on the planet.
And React, although the jury remains out on whether that's a transformative technology or... not.
Working software gets remade because 'delivering' a project, which includes both development and marketing, looks good on resumes, not because the users want it.
If there are quarterly reports there has to be Number Go Up growth.
The net result is enshitification, with income models that rely on exploitation and forced lock-ins, not on voluntary mindshare attracted through quality.
UX-first software that isn't the result of corporate infighting, and isn't automatically aligned with corporate values, still exists in niche boutique developer spaces, but not so much elsewhere.
The biggest issue is that AI changes none of this from the top end, but may change it from the bottom up. If vibe coding becomes more popular bought/subscription options will first be replaced by DIY vibe coded custom tools with some longevity, and eventually by compute soup which improvises tools on request.
I mean real profit, not "adjusted profit."
Making citizenship as hard as possible for acculturated foreigners is a good way to guarantee the culture will collapse from the inside out.
Racism and nationalism are cultural monkey traps. Sometimes you have to let go of the thing you're holding on to to keep it alive.
Meanwhile Sweden has started deporting British citizens who didn't do the paperwork.
If you travel out of a country, your odds of being allowed in without a visa, residency, or a native passport are basically zero.
Holmes is interesting because she's not a freak or a one-off. She's pretty much a generic tech startup CEO billionaire wannabe, finance industry person, and/or familiar political archetype - arguably more extreme than most, but there's nothing truly exotic about the DNA here.
If you look at science fiction AIs from HAL onwards, they usually have limited practical agency - good for handling mundane automation, answering the occasional question, and usurping power, but not in any sense capable of solving original engineering problems.
It's much easier to imagine a big defence-linked computer system like Skynet than something like Claude or ChatGPT, which can improvise working software to a vague spec.
I'm sure a lot of developers have had "I wish I could just tell it to build what I want" moments on the past, but I'm not sure anyone anticipated being able to do it, even five years ago.
So you get permission flags and multilayered data structured with complex interfaces, and there is never any sense of "Wait a minute - we don't need this."
It's much like the way AI art looks meaninglessly busy with too many random lines or other elements, and AI writing is always exactly, precisely, specifically, unnecessarily adverbial.
The best thing about movies is that you get to experience the vicarious thrill of being a heroic revolutionary without having to spill the popcorn.
That isn't even snark, it's plain cold hard on-the-record facts.
My prediction is Musk will become less and less stable, more and more flamboyant and performative, and less and less effective.
He's already saying he suffers from occasional AI psychosis, and there is no chance that's going to get better on its own.
Of all the things we could be doing, why are we doing this?
Is it collective mental illness? Economic hysteria? A cult of virtual personality?
It's simultaneously irrational, obsessive, ecologically and politically catastrophic, and all-consuming - none of which are good things.
This is the first action that has had that effect. Taking out refineries for month after month produced grumbling. Taking out data centres produced an immediate response.
If the data centres didn't matter, Russia and the Trump people would have shrugged and said "Cute..." and/or repeated their standard rhetoric.