Comment by reactordev
2 days ago
Curious how they’ll balance the business needs of moving fast with AI vs quality because my agents aren’t that good. While it works, I’m often having to cleanup afterwards - slowing everything down. I was almost as fast when I had just basic intellisense.
Anyway, I’ll watch the twitch stream from across the pond.
In DJB's paper on software quality he identifies actionable strategies for code quality and code security that were born out of frustration to sendmail's exploit after exploit. Very accessible and fun read: https://cr.yp.to/qmail/qmailsec-20071101.pdf
I would expect this conf to expand on those types of concepts and strategies.
They probably just manage to realize that being seen to be "moving fast with AI" simply isn't a goal unto itself, that it has to deliver something of value beyond itself.
We could be in a tortoise vs. hare situation... Unless we find ourselves back in the conditions of the 2010's again, thoughtfully building software to be high quality and high performance may win out in the long run over "move fast and break things."
Always have been. It’s why the vast majority of disposable corporate garbageware, products chasing a buck, consumer shovelware, etc is built on the shoulders of thoughtfully designed, high quality, mature software that stands the test of time. No popular production software runs on an OS kernel someone vibe coded yesterday. Durable utility is where quality lies, as the cost of quality is able to amortize. Chasing trends is, by definition, costly.
Or at least "value" beyond that reaped by current investors unloading their shares onto "Greater Fool" buyers at high prices.
>Curious how they’ll balance the business needs of moving fast with AI vs quality
Why would they need to do that? Is that even a goal or something that this conference is addressing at all?
There are plenty of alternative software needs that do not need to be AI based nor do they need to change tactics due to the current obsession with AI.
If you know of any that are in need of an engineer, let me know. Every single executive I’ve talked to in the last 4 years is all “How can I use AI with this?”
Well, a couple years ago this stuff all sucked (well, a lot more). Yeah it's in many cases somewhat borderline now, but still - this is frickin magic compared to what I thought was possible just a little while ago.
My question is how far does it go - are the gains going to peter out, or does it keep going or even accelerate? Seems like one of the latter two thus far.
> Curious how they’ll balance the business needs of moving fast with AI vs quality because my agents aren’t that good
I would guess the same way humans do.
Put brain in creative mode, bang out something that works
Put brain in rules compliance mode and tidy everything up.
Then send for code review.
Yeah it's interesting, unless I lean hard on them, AI coding agents will tend to solve problems with a lot of "hedging" by splitting into cases or duplicating code. It is totally fine with infinity special cases and unless you push for it, they will solve most problems with special cases and not generalise or consolidate (gemini, claude code at least both seem to have this behaviour).
I feel like this comes about because it's the optimal strategy for doing robust one-shot "point fixes", but it comes at the cost of long-term codebase heath.
I have noticed this bias towards lots of duplication eventually creates a kind of "ai code soup" that you can only really "fix" or keep working on with AI from that point out.
With the right guidance and hints you can get it to refactor and generalise - and it does it well - but the default style definitely trends to "slop" in my experience so far.
To be fair, a lot of humans also have this problem.
[flagged]
Take a moment to reflect on what you have written and how the casual observer may interpret it.