This is going to kill someone... you can't just make these calls for people. And it's DEFINITELY not a replacement for snowpack tests, etc. at the location you're going to visit.
small downsides like security holes? Those aren't small. Neither is creating a codebase that's an inextensible mess that even LLMs begin to struggle with.
The reality is making good decisions and thinking about approaches take time. AI can absolutely make us faster at it but it's not magic and these speedups come with effort.
This is probably one of the more level headed takes in the comment thread. There's been a concerted marketing push to frame AI maximalism as an inevitability. More or less a "it's going happen anyways so let's go all in".
It's hardly an inevitability though (nothing is... and analogues to the industrial revolution are iffy at best, we haven't ever had an attempted replacement for intelligence itself before).
Society is doing this at an unprecedented cost and it's clear a large portion of the population is uneasy with it. Whether society in the US, Europe, and Asia will continue to allow such investment at the expense of everything else remains to be seen.
A part of the puzzle that rarely gets discussed is something that predated LLMs entirely-- "software engineering" and "programming" have been conflated for a long time now and there's a huge gamut of roles out there.
The practice of writing code, or programming, in recent years has really fallen into two buckets:
The vast majority of folks are given a task, they write code to complete that task, and the task completion then counts towards some objective (eg; a new feature, product or fixing a bug). Perjoratively, they've been known as "ticket takers".
A much smaller group have instead worked in the other direction-- identifying where improvements can be made to a product, piece of infrastructure, or pain point and transformed that into tasks that can then be solved via code.
How much of a role you play in that strategy and formulation has been the real differentiator. Not so much what you know. While these are correlated, they're very different.
At a high level, it's been the difference between "developer" and "engineer" but the reality is the titles have become somewhat meaningless in recent years where many "engineers" are just doing the same CRUD tasks over and over.
The reason this matters is that at some point, you can only abstract so far... the requirements for what to build have to come from somewhere. At the most extreme case, there's only the CEO and a company that's nothing but AI agents. In the least extreme case (today) each line worker could manage 1 or more LLMs/agents.
It's not entirely clear to me or frankly a large portion of those in the industry that we're suddenly on pace for one outcome vs another. But I do think that software isn't particularly unique here other than it was an initial starting point for LLMs to deliver value. All white collar work is at risk including CEOs.
And if that happens it would be outlandish to think a utopia emerges... the opposite is far more likely.
It's not entirely clear to me that the opposing argument is well-formed either. You constantly see numbers and statistics being wildly mis-used or overextrapolated.
It's not just that there's a circular deal it's that they're prevalent. And worse, with frontier labs IPOing seeking astronomical valuations that means a lot of the public is now exposed too (even if they don't all get fast-tracked into eg; the SP500).
The problem is the valuations assume astronomical growth... that is likely impossible for all of them to simultaneously achieve. Which means something's got to give.
> the big providers are charging full freight for inference
They're not and it's not clear why you seem to believe that. The immense capex for buildouts, training costs, etc. are not rolled into inference costs. Moreover, companies are already rapidly starting to re-evaluate token spend.
Generally, I agree because what happens is the messaging around AI is doing more, faster. Not using AI to deliver at a higher quality level, etc. But I think it boils down to incentives and discipline. So given the incentives we have today at most workplaces faster AI will just be used to produce more slop.
Right. Which means tokens are actually being priced well under cost once you factor in all this datacenter/GPU capex. Also worth noting the datacenters are not purely for training. They're for inference too.
Honestly... not that dramatically. Each release is much more marginal. And quoted official benchmarks doesn't translate very well into the real world.
4.7 regressed hard in some ways. But a compounding factor too is that the claude code harness seems to nerf the model after a few months. Probably to reduce token use.
So far 4.8 seems less verbose but we'll see in practice what it translates into meaningfully.
Having tried something similar, the perceived speedup does not, in the steady state, last.
To get a quality, lasting, result you're ultimately having to carefully study everything otherwise you end up quickly accumulating cognitive debt and the speedup soon shrinks as you're constantly having to revisit the initial approaches.
Yep. And this is why as hard as AI companies are pushing that these tools can be a replacement for expertise, it's ironic that the experts are the ones that often get the highest ROI because they know how to converse about the relevant subjects with a high degree of precision (and know what to look for, what to challenge, etc.).
Yep. Trust is easy to lose, hard to earn. A nondeterministic black box that is likely buggy, will almost certainly change, and has a likelihood of getting enshittified is not a very good value proposition to build on top of or invest in.
Increasingly, we're also seeing the moat shrink somewhat. Frontier models are converging in performance (and I bet even Mythos will get matched) and harnesses are improving too across the board (OpenCode and Codex for example).
I get why they're trying to do that (a perception of a moat bloats the IPO price) but I have little faith there's any real moat at all (especially as competitors are still flush with cash).
It's quite prevalent in tech too-- however, folks tend to be quiet because the "use AI for everything or else" hammer is being used across the industry.