I think that's a pretty bold claim, that it'd be different every time. I'd think the output would converge on a small set of functionally equivalent designs, given sufficiently rigorous requirements.
And even a human engineer might not solve a problem the same way twice in a row, based on changes in recent inspirations or tech obsessions. What's the difference, as long as it passes review and does the job?
One could argue that's a cynically accurate definition of most iterative development anyway.
But I don't know that I accept the core assertion. If the engineer is screening the output and using the LLM to generate tests, chances are pretty good it's not going to be worse than human-generated tech debt. If there's more accumulated, it's because there's more output in general.
Only if you accept the premise that the code generated by LLMs is identical to the developer's output in quality, just higher in volume. In my lived professional experience, that's not the case.
It seems to me that prompting agents and reviewing the output just doesn't.... trigger the same neural pathways for people? I constantly see people submit agent generated code with mistakes they would have never made themselves when "handwriting" code.
Until now, the average PR had one author and a couple reviewers. From now on, most PRs will have no authors and only reviewers. We simply have no data about how this will impact both code quality AND people's cognitive abilities over time. If my intuition is correct, it will affect both negatively over time. It remains to be seen. It's definitely not something that the AI hyperenthusiasts think at all about.
I stated plainly: "we have no data about this". Vibes is all we have.
It's not just me though. Loads of people subjectively perceiving a decrease in quality of engineering when relying on agents. You'll find thousands of examples on this site alone.
Turn off swipe typing in the keyboard settings, and tap typing works a bazillion times better. It's like toggling a completely different codebase behind the keyboard and resetting it back to when blind-tap and autocorrect actually functioned as expected.
I assume the code that checks for tap vs start-of-swipe is to blame. I have no idea why that would cause word recognition and/or autocorrect to work so differently, but it seemingly does.
I've noticed the iOS keyboard has fundamentally different tap recognition based on whether swipe typing is enabled.
It looks the same but behaves differently enough that I have a hard time believing it shares code. When I turn off swipe, my tap accuracy goes MASSIVELY up, and a lot of the autocorrect screwiness seems to abate considerably. I can go back to blind thumb typing.
That said, swipe is so useful, I’ve left it on, and I deal with the degraded tap behavior. But maybe that’s a trade-off for you to consider.
I can't believe this is it. But this is it.
Too bad there's no quick toggle to turn it back on?
It's possible to create a shortcut for it maybe. I currently have a back tap bring up a menu of different shortcuts I use. Shortcuts is another aspect that's really under utilized because the UX just sucks so much.
I’m pretty sure temp chat mode doesn’t prevent the model from accessing your past chats and personalization. It just means that chat won’t be saved to them, to be seen in the future. It’s the same as incognito mode in browsers—it doesn’t prevent your search history from being used; it just keeps that session out of it.
If the experiment had been based on the idea that that option isolated the question, it may have been flawed. I found my ChatGPT’s o3’s accuracy went way down when I cleared personalization and deleted all past chats (turning off extended memory would’ve been equivalent, I think).
Importantly, only once did the o3 reasoning mention it was fishing from my past chats—that’s what clued me in I messed up the isolation—but the guess rate was still radically different from all the times before once I cleaned house. That suggests to me that it was quietly looking before, and it just didn’t make the cut for explicitly saying so.
The fact you had to clarify "at dancing" is why they're different.
Think about it as a decomposition where "dancer" means "dancing person."
In the simple case, both "blonde" and "dancing" separately modify "person." If you diagram that it’s a Y: either modifier could be removed without changing the meaning of the other, and their order isn’t important.
In the complex case, "bad" modifies "dancing," which together modify "person." That’s an ordered chain, which is more complex to build and comprehend. Your clarification illustrates the chaining and why it’d be a fundamentally different meaning if that wasn’t understood.
I’m not even touching whatever you were going for with the blonde/brunette thing. It's plain they used the example because there's no possible way hair color could be a modifier for "dancing," and they wanted something unambiguous.
You can also assign it to the triple click shortcut in Accessibility. You probably can to the double/triple back taps too, though I haven’t tried.
I do use a standalone Lectrofan for sleep as I prefer my noise machine to be across the room and Alexa-controlled (via a smart switch), plus it’s louder and the brown noise is “browner.”
But I keep iOS BG sound mapped to the triple-click shortcut for when noise-cancelling just isn’t enough in loud restaurants etc. It works great with AirPods for reducing my noise sensitivity issues.
This is technically the law in California, too (and possibly that’s where you live) but in practice the ability to do common law name-change has been abolished at a bureaucratic level.
Since at some point everything official now routes back to either a passport, an immigration/naturalization document, a court-issued name change doc, or a marriage certificate, those are effectively the only ways your name can be changed.
The problem is that all the law can do is make it legal for you to use whatever without it being considered fraudulent to try. But if your law is like CA’s it doesn’t specify other institutions other than possibly state government ones have to honor that and, in particular, it can’t constrain the federal government at all.
So that leaves the DMV as the one possibly effective way to do common law change, on the off chance somewhere will just accept your license as proof of identity. But now that driver’s licenses are subordinate to passport info or equivalent via Real ID, that route is pretty much toast too.
You might still be able to get the alternative state-only DL / state ID with a common law name, and maybe open a bank account with that, but then the credit reporting companies (or Chexsystems) don’t have to honor it so you’re possibly screwed anyway. Plus without a Real ID you’ll have to show a passport to fly domestically, and that will have the name you don’t like.
And, of course, none of this helps with your paycheck because you can’t satisfy an I-9 with a DL. It requires a federal document, too, which—if your state info doesn’t match your federal info—needs to be a fully identifying federal passport/equivalent. So even if you get the bank account with your chosen name, you might run into issues with your checks being to a different one.
At this point, it’s just not worth trying common law name change anymore. You either flip a few hundred for the official change or you accept the fact that you’ll have a public name and a private name.
(And I say this as a “Geo” who strongly dislikes seeing “George Jr” on stuff so I feel your pain. I just tell my employers that my given name only goes on paychecks, benefits, and tax forms, and is to never be used publicly. That has always worked.)
And even a human engineer might not solve a problem the same way twice in a row, based on changes in recent inspirations or tech obsessions. What's the difference, as long as it passes review and does the job?