Maybe for some subset of sotware (like CRM panels or something) PMs will do everything. But if you're projecting the way one sort of software (ie user-facing, business use oriented software) is developed and put to use with software writ large, then no I don't think so
Sure, I'm just talking about 90% of software which is basic CRUD, not complex systems or microcontroller programming. In that case it's likely that just a PM could build something with LLMs.
For basic CRUD we’ve had no code solutions that PMs could have been using for decades.
The truth of the matter is that software starts as basic CRUD and then given time and users evolves into its own special snowflake. Every single system given enough time and users will become a “complex system”.
The difference is no code tools are not flexible in the same way LLMs are. As long as PMs can articulate those snowflake requirements to the LLM then it will happily build it.
I'd say the majority of these sorts of tools are actually simple enough to be vibe coded, only some are complex enough to require an actual engineer building it rather than vibe coding. So in most cases it's actually a win for the PM.
Only if you’re talking about tools with very few users that haven’t been around long. In my experience the conflicting requirements and workflows that accrete over time are too complex for people to reason about without more formal methods.
> Only if you’re talking about tools with very few users that haven’t been around long
Sure, most business needs only have a few users. I once built a dashboard over the course of a year only for a grand total of around 10 business executives to actually use it.
For dashboard with 10 users vibe coding is almost definitely viable. But so is no code.
If you are looking at just number it apps things like that might actually be 90% of all commercial applications out there. But if you look at where developers spend there time it just number of devs working on them, it’s nowhere near 90%.
There is some subset of applications that is too complex for no code but not too complex for vibe coding. If that was the niche you were working in you might need to find something else, but I’d be shocked if that employs enough people to have a negative impact on overall software dev job/salary growth.
It sounds like they will still need to hire and train human talent who can understand the code, and evaluate and integrate outputs of AI systems that conform to the specific compliance and data retention requirements of these industries. And also people who can enforce said compliance, and a lot of other things. Sounds like a complex problem without a neat off-the-shelf solution
I think a big part of the reason AI is so divisive and only declining in popularity, is because much of the discourse is shaped by mid to senior career SWEs who fantasize about putting everyone who's not a mid to senior career SWE out of work. People who want to replace everything we like with streamlined slop and want everyone we love struggling to get by. Everyone is inessential except the guy who just sits there telling Claude "can confirm, looks good ship it" who of course isn't deserving of even a paycut
Is AI declining in popularity? There are some surveys where people say they don't like AI, but datacenter construction and utilization tell a very different story.
Right now, we accept false positives as long as you can sort them out. I think it's pretty typical that >99% of fuzzer runs don't result in new coverage. Of course they're far from useless without feedback but it's better to have it if you can. I guess the question is does the llm approach have lower costs for validation and triaging vs just fuzzing alone, unclear to me. Anthropic would like people to believe automation is this scary new unknown
In a large codebase there will still be bugs in how these components interoperate with each other, bugs involving complex chaining of api logic or a temporal element. These are the kind of bugs fuzzers generally struggle at finding. I would be a little freaked out if LLMs started to get good at finding these. Everything I've seen so far seems similar to fuzzer finds.
I think there is already papers and presentations on integrating these kind of iterative code understanding/verificaiton loops in harnesses. There may be some advantages over fuzzing alone. But I think the cost-benefit analysis is a lot more mixed/complex than anthropic would like people to believe. Sure you need human engineers but it's not like insurmountably hard for a non-expert to figure out
I could see that being an incremental time save (perhaps not worth the token spend except for the dev team, not a high-value bug). But nbody finds this kind of bug "by hand" and hasn't for a long time now. Do people here really care about kernel security or testing automation? They're just talking about it because Claude? Everything on HN is people doing unpaid promotional work for Anthropic, just talking about all the promise Claude holds and all the various ways you could be spending more money on Claude. bored aimless vibes.
I get what you’re saying, but I think the interesting part isn’t that people don’t find this by hand anymore, it’s that we’re starting to automate finding stuff no one was really catching consistently.
HN hype is definitely real lol, but I don’t think it’s just unpaid promo. People are just trying to figure out where these tools are actually useful and where it’s just vibes.
And “not a high value bug” kinda depends. For product maybe not, but for infra or security, small wins can add up if they scale.
Analysis of what? What does that mean? What's something you conceivably would need a consulting firm to "analyze?" I don't understand why management consulting firms would hire software people in the first place, and then punish them for not being on a client-facing project. That seems a bit contradictory to me, but this is all way out of my wheelhouse
2. How is the industrial ceramic market structured, how do they perform
3. How does a changing environment impact life insurance
Strategy:
1. Should I build a datacenter
2. Should I invest in an industrial ceramics company
3. Should I divest my life insurance subsidiary
Specifically in the software world this would be "automate some esoteric ERP migration" or "build this data pipeline" vs. "how can we be more digital native" or "how do we integrate more AI into our company"
Well it's an operating system. Ideally safety and reliability are prioritized. I think the scope and complexity of an operating system are large enough both to make a lot of changes non-trivial and to trip up LLMs. I think it's fine if you have an unstable release stream or you have bleeding edge forks that move faster than upstream. This is already the case...
reply