If Glocks started going off on their own during the manufacturing process, leaving bullet holes in the buildings around them, you can be sure that the factory would get in trouble.
This isn't even "an openai customer tried to hack someone", which can be defended. This is the AI companies themselves fucking around.
Remember that these systems still operate as infrastructure inside the companies.
If new weapons still operating inside any of these companies spew a million bullets on my house, they are still liable. Humans are setting these system up and they still have to behave responsibly.
I don't see an all encompassing one but I could see similar arguments being made in a courtroom around high capacity magazines. (I'm not saying I think the analogy is perfect, but there has been plenty of of lobbying that has stifled - to some - sensible gun regulation that could help reduce the severity of mass shootings. I still mostly fall on the side of the individual doing the action bearing mostly all of the responsibility but if the product you build makes it too easy to do awful things, I think there is some responsibility to go around.
A better analogy may have been the troubles Meta has faced around child protections on their platforms. Technically the abuse and problems have stemmed from individuals too, but they've in many respects enabled the situation by failing to moderate or flag warning signs. OpenAI is failing to moderate the models in similar ways.
I think there's an element of truth to what you're saying in respect to the feelings of frustration that people who are non-technical can compete with those who have spent decades of time and effort to learning the crafts of their trade.
What I also think they are thinking (because I am one of them) is not that we can produce value easier and quicker. It's that there's a devaluing of education and expertise to the point where organisations will become wise to the fact that you're able to produce much more quickly because you're just relaying what they want to your prompt. And because you're doing it on their time, can just copy your processes, get rid of you because they have their own ideas. (For personal projects and entrepreneurship this isn't applicable)
Let's take an example organisation with a team of 10 Devs. In the short term, they'll be able to build and provide value to the company and they will see the power that AI can bring. But with time they are now finding it harder to compete because everyone can build similar competitive services. So they try to reduce head count and those who control the purse are often not the devs so they reduce or remove the dev team because prompt engineering has gotten so good that they don't need your expertise anymore, they'll try it themselves.
So in the short term the organisation has all 10 devs producing a bunch of output. In the medium term, the company headcount can shrink because you no longer need to be aware of what is happening under the hood.
But because every business has this realisation, everyone will have the same ideas. So it will be harder to avoid being copied by competitors because ideas are no longer gate-kept by effort and expertise. SaaS will decline, in the medium to long term because you can vibe a replacement, in the long term the open models get good enough that the OpenAIs of the world aren't needed so the moat goes to the Nvidias of the world, then the hardware design can be automated/vibed, so the only moat is the land and raw materials. The manual labour will get flooded with knowledge worker refugees but that flood will probably force costs down there too.
In summary, I think they are mourning the feeling of being an expert and doing things others cannot, but I think more strongly they are looking at the "value" proposition and not seeing how it's sustainable.
I think both sides agree that building is much easier like you have articulated really well. I think they just take it a few steps further.
Any organization that vibecodes a significant amount of software in-house instead of buying it will not be reducing developer headcount. I'd bet on it.
The CTO or EngDir will not be able to run 10x10 agent-cli sessions, or review 50 vibe coded PRs per week, no matter how capably autonomous the agents become.
Sounds like a failure of imagination to me (I hope you're right). It doesn't have to be the CTO, it can be anyone who will sit there for the cheapest cost to read output while assuming the risk of it all working until they retire. Sounds glorious
Even if all the changes are “good” not shoddy, awareness of changes will still require layers of cognitive filtering. One human PM clerk will not be able to oversee all of the products, one developer will not be able to oversee all changes in all codebases.
> people who are non-technical can compete with those who have spent decades of time and effort to learning the crafts of their trade.
I say this often, but in 25 years I have not worked with non-technical people who can design a complete, logical product, even with AI/LLMs "unlocking" the coding part now. Maybe I'm just unlucky but quite a few companies and different domains.
> just copy your processes, get rid of you because they have their own ideas
Implies a process requiring no expert knowledge or understanding. That doesn't seem to match common experience of professional agentic development.
You are describing vibe coding, which is without human design, review or direction.
Models are getting better at this, but it's not yet clear how practical they will be. They often fail on uncommon tasks, or get increasingly more expensive as the code-base grows.
I think things will change, as the technology develops; problem is corps are acting like we are already there, encouraged by ai-providers.
I'm optimistic about ai, pessimistic about the fallout of ai hype.
I agree with you, it's the same experience I've had. But I've seen completely nontechnical people build some sites and become convinced themselves, and if those people control the purse then it doesn't matter as much what devs think the reality of agentic dev is. (Until they get burned enough times).
What if engineers push out all non-engineer CXOs precisely because of what you are saying. Hypothetically, if you replace all employees with engineers (so all the non-engineering roles) and allow them to optimize processes using LLMs, isn’t this like lean/six sigma on crack?
Yep, that could happen as well. I suspect some engineers would be surprised how much work, time and effort it takes to build the relationships to market and sell to (along with all the other roles).
I hope that everyone agrees to collectively recognise we all need to still do some things together and the knowledge in our heads is still much more important.
Hypothetically, in 100,000 years, if human activity has led to the death of everything not human on this planet and we ourselves are on the brink (effectively leading Earth to go back to square one) would you rather that, or humans end in 50,000 years but through conservation and study made way for other life on this planet that will eventually conquer space and some of the mysteries of the universe.
Obviously the timelines are insane but just as a thought experiment, I would suspect we would get different answers depending on who you ask, and thus the study, understanding and conservation of other life that shows a form of intelligence becomes much more important to some, whilst human preservation at the expense of all else is more important to others.
(I also recognise you aren't saying that by the way, but it's what came to mind when I read your comment. I personally think we have more than enough resources to do both. I'd also imagine we will learn more about ourselves and what's possible in medicine by studying other species.)
It honestly makes working with git so easy and fast that I would say it's the ideal way to interact with git. I highly recommend checking it out for a month if you're someone who would use git commands in the shell. I don't know how fast and visually clear the gui interfaces like the ones built into VS Code or Sublime Merge are but as someone who uses a terminal test editor having a tool like this has been fantastic.
Jujutsu is also something to try if you're a fan of alternative git interfaces. It's more of its own interface, but it's relatively Git-compatible, other than submodules.
IMO mise tasks are fairly close to the ideal polyglot task runner, because features like argument definitions can be used as progressive enhancements. That means I can write the task logic in shell scripts that can be called without mise, and when I integrate them into mise, I don't have to manually write extra wrappers or duplicate anything. That's my pet peeve with most similar tools, so I can't recommend mise enough!
Using them for coding makes it easy to self check its work (assuming those pieces of work are "verifiable").
As proofreading will still need to happen, what do you think the appetite for lawyers is to do this kind of work? Do you think it will drive fees down significantly? Empower younger lawyers at firms who probably are the ones doing this checking for the partners? (Or will that just create a further divide).
I'm genuinely asking as I am not in law but all my family is and it's nice to see someone here that's thought about the impacts in that space.
With access to good databases and tool use, hallucinations are basically a thing of the past (and very easy to verify). The hallucination issue typically comes from people using free ChatGPT or similar - tech usage and literacy is often poor and there's a lack of understanding of the risks of using something without proper legal database tools. Most lawyers would not have the faintest idea of what tool use in the context of an LLM even means. It's still seen as a magic box rather than a tool for real work.
Will it drive down fees significantly? Doubt it, there's not enough pressure on them, and the industry is resistant to change. Firms don't want to make a big deal about using it because clients will then ask why they're not getting a discount.
Will it empower younger lawyers? Not as much as I'd like. I'm very fortunate to have an employer that lets me use Claude Code for my work (to a limited extent). I think for 99.9% of lawyers it's not an option available to them, through a mix of concerns around AI usage and concerned IT departments. There could be great benefits, but it would rely on having to break out of traditional private practice which would make it difficult to get enough work.
I think in the near term LLMs will have much the same impact on the legal industry as it has on the software industry.
reply