At least in my work, this is sort of like asking "If you don't enjoy CI/CD or the cloud, why not do without it?" It's becoming integrated into every process at this point.

And many employers now require you to code faster, which is only possible with AI tooling. They don't understand that coding faster isn't always better.

Is this related to what business your employer is in? In other words, is their business producing code, or is code written to support some other business?

Or is this just everywhere now?

Everywhere. Normally adding an extra endpoint to the REST API would take a sprint. Now PM expects you to do two. Only way to get that done is a lot of vibe-coding and delivering sub-par results.

I am amazed that anyone ever needed a "sprint" to "add an endpoint".

It usually speaks to the approach you are taking, but I also believe sprint-based planning leads development teams to make work fit into it instead of needing natural time to do it.

Let's remember that all of have probably built full systems over a weekend. What organizational dysfunction has led to us needing a sprint is what needs fixing.

How is that enforced though? At any normal workplace they will ask you "how long is this going to take"?

It depends on the management. Mine asks that, but others within my company get "this is going to be done by [DATE]".

And no one argues about unrealistic deadlines?

Of course they do. Generally the end result from this is bad - reduced in scope, buggy, and/or unstable. But that's an issue with communication/expectations/management. It'll keep happening either because of politics or the fact that a sub-par deliverable still meets the needs of the overall organization. Until it doesn't.

A few firings of those that do under this job market convince the rest not to.

Job market in Western Europe isn't doing so well right now. Better not make a big deal out of it.

"How will you be leveraging AI for this task? Why can't the AI do it faster than that?"

These managers literally believe, or are made to pretend, that AI is better than you and faster than you. Any argument to the contrary is a CLM.

Ah yes, the free labor market, where for some reason we’ve given one side a monopoly on whether you can see a doctor without going bankrupt.

My complaint with anthropic is actually the opposite. They seem too focused on building this suite of products (because they want lock-in), but they can’t even get the availability and speed on their models in an acceptable state. It really does seem like they’re falling behind google and OpenAI at the moment.

This is really SDE pay coming back down to earth. You're still making way more than even most white collar jobs at a startup.

If there is, it's not listed as the solution on the puzzle website

It's impossible to completely evaluate a product before buying. That's why it's important to shame once you have enough experience with it

> A reasoning error has an infinite, unpredictable blast radius.

Says who? It’s quite easy to limit the blast radius of a reasoning error.

In 2024, a Chevy dealership deployed an AI chatbot that confidently agreed to sell a customer a 2024 Chevy Tahoe for $1. It executed a catastrophic business failure simply because it didn't know the logic was wrong.

Sure, you can patch that specific case with guardrails, but how many unpredictable edge cases are you going to cover? It only takes a user with a bit of ingenuity to circumvent them. There are already several examples of AI agents getting stuck in infinite loops, burning through massive API bills while achieving absolutely nothing.

You can contain a system failure, but you cannot contain a logic failure if the system doesn't know the logic is wrong.

This would be more convincing if a single car had been exchanged for $1.

It didn't happen. Seems the bug was "contained".

Sort of undermines your point re "catastrophic business failure" don't you think?

You literally can contain a logic failure. If I execute logic on my computer that’s not connected to the internet it can’t get out of the box. Done. Contained

> but how many unpredictable edge cases are you going to cover?

This is the wrong question. The correct question is what specific subsets of cases do you allow, similar to any security question

How so?

Suppose you had:

Math() Add() Subtract()

Program() Math(“calculate rate”)

This is intentionally written vaguely. How do you limit that these implementations ensure Program() runs and does the right thing when there is no guarantee Math() or its components are correct?

Normally you could use a typed programming language, unit tests, etc, but if LLM is the ultimate abstraction programs will be written line above. At some point traditional software engineering principles will need to apply.

> The "song of the summer," the latest TikTok trend

These are 2 cases where you absolutely need a personality to go along with the song. Department store ads are probably already AI.

> If we assume that we fix bugs faster than we introduce new ones and we assume that the AI tools can improve further, the question is then more how much more they can improve and for how long that improvement can go on.

Why would we assume this? Historically it's obviously inaccurate since the world began with 0 software bugs.

There's another problem

Who guarantees that those AI fixes don't introduce new bugs?

do you... honestly not believe that system design is real???

That's not what I said. I said system design is not the exclusive domain of humans, so anyone thinking they possess some special knowledge of system design that an LLM isn't capable of obtaining are fooling themselves.