Hell, the fact we don't know what consciousness is means that some LLM systems (or trees! or planets!) might already have it, and we'd be none the wiser. Not that I believe this, but as these systems get larger and more sophisticated it's an increasingly concerning proposition that we have no way to validate either way.

Nah. C-suite has (a) money and (b) legal and fiscal obligation. You literally need a human being certain legal operations, like incorporation, salary, and lawsuits.

I wouldn't count on lawyers being deemed necessary by the law for as many actions as they have been so far.

Maybe. Or maybe the C-suite salary will be cut to $50K/yr. That would make them marginally more expensive than AI, but for that price the shareholders get an extra layer of insulation should things go awry (and they won't be held personally responsible for corporate malfeasance).

> I wouldn't count on lawyers being deemed necessary by the law for as many actions as they have been so far.

Maybe, but that's pretty far down the road and would require restructuring the whole jurisprudence system.

I know the software engineering world like to think the C-Suite are disposable, but that is only because you really have no idea what they really do.

The death of SaaS could also be a thing in the (not so distant) future.

Just because it's faster and cheaper and requires less skill, doesn't mean it's instant and free and requires no skill.

I think we’ve barely scratched the surface of what software can do if it’s completely custom built, specifically for individuals or small businesses who could’ve never afforded it, pre-ai. Home automation, health and disease management, all sorts of administration, conversational ai over your life’s data - most people have no clue what’s possible once it’s custom built. With close attention, good design, creativity, and without giving away privacy or independence.

In my limited spare time I’m building up this tapestry of custom services for myself and my loved ones, and having a great time doing it, while trying to learn how I can offer this as a service.

Judging by the reactions I get when showing people what I can do with my phone, and no that’s not available on Google play, I think the demand will be there.

Not to say you shouldn’t! But it’s worth being realistic about income compared to (current) software salaries.

reading HN these days feels like constant gaslighting. this is so far from my reality that i wonder what folks who are saying this are working on

I don't use AI that much. When I do, it's often surprisingly competent in the topics I'm a specialist in, and it can do a lot of good work quickly. But other people in other organizations still continue building upon my work instead of asking AI directly. Apparently because they don't know what to ask.

Nobody knows how long this will last. Hopefully long enough that I can afford to retire, because I'm old enough that I would face ageism if I had to find a different job.

And if we get there, we will probably have a serious political debate on how many digits should there be in the wealth tax rate.

I just had fable 5.2 (high effort) review a change. It hallucinated 2 problems that I was only able to correct because I also understand the code.

LLMs can do any job _in theory_. It’s not a very interesting argument. Their effectiveness at various jobs is much more interesting.

and the product, the team, and (no matter how shallowly) the organizational pressures and politics. Whether we like it or not, those are part of the problems developers solve. Even something as rudimentary as knowing what is feasible to test given your deadline isn't something you can encode in a model[1]. Code does not exist in a vacuum.

I say this even with the latest models, it seems as soon as your development practice falls out of the ideals written up in blogposts of the previous ZIRP decade, you have to course correct. I use this term because the AI is technically correct (the best kind of correct) but practically useless, if not also irresponsible.

Recent anecdote (literally today): we've been having Claude write exhaustive test suites to eat at the coverage backlog. Lately, we noticed that tests in develop are even more flakier than usual. Our whole suite runs in 30-40min but lately it could exceed an hour and then fail.

It seems Claude took "exhaustive tests" to mean "exhaust the compute resources of the Jenkins node". What's worse is it wrote those tests for features that have been disabled and would soon be rewritten anyway.[1] I've always said that LLMs are great for generating test suites and this is the first time an LLM "failed" this assignment for me.

Code is cheap. The right code (or lack thereof) for the right situation takes meatspacebrainpower.

[1] The solution here is to run your company/team 100% as agents so they have the full context of project timelines, resources, customer feedback, etc. ala stevey and gtan but my reasons why I don't believe in that has already been articulated elsewhere.

Then who directs the swarm to create the most value? An executive can’t. They have other things to do so they delegate. That person will be an engineer and their attention is finite too so they will delegate. To other engineers.

Here’s the best part. You write back and say I’m wrong. Guess what? That’s engineering too. Some people have to sort out all this nonsense and the stakes are higher than they’ve ever been and now engineers are supposedly not relevant.

But yeah when you get everything under control, what’s the day to day look like? Pretty freaking cushy. And then it relies on intelligence to figure out what next. Clients always want things done and having those conversations won’t get much easier with AI. Keeping up to date. Jesus I’ve exhausted myself writing all this. So much work!!

A readable link for those of us without Twitter/X account

The problem is that software in corporations has typically been product managers at the front, software engineers in the middle, and QA at the end. Software developers who also have good product sense and do diligent QA are still needed. Programmers who treated their job as leetcode exercises are not needed anymore.

If that's the end user, they're not happy. End users generally don't like doing unpaid involuntary QA work.

And then also it's as you say, customers do need help from domain experts (whether they admit it or not).

The sheer amount of time that gets wasted when people who are awful at expressing what they want just slap together a ticket with little understanding of what they're asking, what the software does, or what the customer wants or needs. Which is far, far too many of them.

Previously, eventually I or another person who does know that would get to it, cross out all the utterly irrelevant parts, strip out the parts that were clearly a completely unrelated ticket, point out to them that they'd asked for the font to be coloured both blue and red simultaneously (neither of which matched the style guide, the company or product colour scheme, or any other part of the UI) and so on and so on... it might takes weeks from start to end for a ticket to actually get expressed well.

Now, the LLM reviews it, adds a whole lot of questions, points out a bunch of nonsense, and sends it back to them. Even in cases where the LLM makes some mistakes, it's 100% fine because the human is forced to think about what they want. The LLM is great at natural language, and it doesn't matter what mistakes the LLM makes.

The quality of the software is going up and up, purely by finally getting a handle on awful requirement and ticket writing.

The human isn't replaced at the start of your sandwich, but now the LLM makes that human do it properly, at speed. Massive, massive reduction in time wasted and significant improvement in software quality.

  {(P(AI mistake) * P(human spots the mistake | AI mistake) * P(human fixes mistake | given AI mistake, human spots it) - P(AI mistake)} * cost of AI mistake

So even if you can “make things right” or “tell the difference” as you say, the value of any given engineer is dropping.

Not remotely the case. LLMs make mistakes very frequently, much more frequently than humans do.

My last project had 2 guys from India and one guy from a consulting company that I take any LLM over them. The guy from the consulting company spent more than a month trying to implement a design that actually was not implementable. I obviously also blame the supposed professional design company that produced the design, and I mean it took me about 6 hours spread over a day and a half to figure out waitaminute this is actually not doable! But I have to wonder how badly would an LLM have done. I sort of think the same, because probably an LLM is never going to announce - hey, this design as stated is impossible, some parameter needs to be changed.

human mistake, AI spots the mistake

Human mistake, human spots mistake

AI mistake, AI spots the mistake.

Everybody is always talking as if they don't make mistakes, but every software engineer does; it used to be that human caught those mistakes and fixed them, but if AI catches those mistakes and fixes them, then that is also a decrease in value of human.

I also don't buy the idea that there is special human stardust that gives us permanent irreplaceable economic value. Even if the tacit knowledge that the current generation of engineers were able to build up in the years prior to ~2025 does give us value that won't be completely replaced for a long time, why would you assume that the delta between what you can do and what claude can do will still be worth $200K in 5 years?

This is the real fear. We had a flood of engineers who entered the profession in the past 10-15 years who were decidedly average. No real interest in software development beyond "it's a cushy desk job, or better yet WFH job, that pays really well." No interest in whatever product was being developed by their company. Just show up, their PM assigns them a ticket to make a widget with a button, they use their TypeScript bootcamp skills to code up a widget with a button, ticket moved to In Review, job's done.

Now an agent can do that, and you need fewer of these types of people. Like, a lot fewer. Every engineer on your team who isn't a 10x-cum-100x engineer feels like a massive lost opportunity.

The "10x" engineers I know have never been more excited about their jobs. I'm really unsure what's going to happen to the average ones.

I feel like that’s where the average ones will go.

Clearly there is still some work to do, but the work I do now is very different, and for me, much less enjoyable.

I'm not solving challenging logic or algorithmic problems anymore, but basically just iterating on features with agents or getting them to look into things for me.

It sucks because I used to be very good at these things, but my edge against less competent SWEs doesn't really exist anymore.

More annoyingly, people I work with are producing a lot of code that I think the design of kinda sucks, but agents can force the design to work and so it works. I suggest we do things another way, but I'm now just slowing things down and it's hard to argue against the comment, "if that becomes a problem we can just get Claude to refactor later".

I guess what I'm saying is that I think think there's a really crappy job that will still exist for a while around managing and contributing to these agent-generated codebases. But skill is less about your technical abilities, but how good you are at hooking up agents to things and getting them to solve problems. This doesn't interest me at all though. My desire to spend a lot of time design and thinking about problems, and deeply understanding the codebase I'm maintaining now just work against me.

I think there will be SWEs by some definition, but the job as it existed a few years ago is gone. Not even saying that's a bad thing, but for someone like myself, it may as well be dead.

It's happened before and it'll happen again. Some examples from the 20th century...

Big companies used to have printing departments to which you'd send a print job, walk over and pick it up, nicely stapled. Now the printer sits there 24/7 awaiting your print job. You have to staple it yourself though.

Big companies used to have secretaries & clerical staff who'd type up your letter and send it off for you. Now we have keyboards and email. You have to send the letter yourself though.

Big companies used to have "code typists" who typed out huge amounts of code. Now they have agent managers doing the typing. Getting good output is a challenge, but asking for the output you need is so much better than RSI from typing and so much more effective than debugging based only on one's puny human knowledge, that we're suffering through. So you have to assure the quality of the work yourself.

It's strange to see people take to their blogs and loudly declare themselves unemployable. I doubt anyone'd hire this author and that'll make their prophecy come true - for them.