rplnt

I use this pattern in mysql. One row, and one update query gets you a block of IDs. Don't want to attempt to write it out here without looking, but the trick is using the last insert id in the query (and on the result).

It's one query, so shouldn't require a plugin I hope.

Use snowflake? It leaks some info, but not nearly as much.

That's why you wouldn't run any of these outside of some sort of sandbox, right? I found that Opus 5 is finally aware from the go that it is started in a specific sandbox and is able to either ask me to enable it somehow or hand off commands for me to run. That's a major improvement over the classic "can't seem to be able to install playwright, let me implement my own browser quickly" when asked to increase font size or something.

One of the recent features I love the most is that they virtually the same color for symbol highlight and text selection. No visual feedback on double clicking a symbol to see if it is highlighted or not (and it sometimes isn't, because they managed to break that too!). Cannot highlight part of the word that happens to be a symbol either. What a lovely feature. It shows that they are not using the product much anymore.

> dishonest

This is what pissed me off the most. Make it slower, rate limit it, move the credits to other time slots, idk.. but returning BAD results? That's the worst approach you could take.

Searching for webpages, at least on google, have died like 10 years ago. Even if you knew the exact domain and words, it wouldn't be able to find it somehow. Always pushing "fresh results", changing your words (quotes or not), translating languages, etc... I tried to be smart and ruined everything that was good about it. On top of that web changed a lot and loads of content moved to private platforms. And now "AI".

> I personally think people sense nerfs more often than they happen and that it's often about honeymoon effects.

I believe in temporary nerfs. Operators reducing quality significantly to increase throughput for whatever reason (high demand?). Same session, model being completly incapable, and it being back to normal the next day. Experienced that with Anthropic models way too many times. Never on weekends, usually during US work days.

It's been a few months since I last recall this though, must have been pre-opus-5. And I know there are benchmarks for this as well, hence "believe".

The project description said this

> Song: as far as I could find, it comes from this YouTube video from 2024

and I did not investigate further

One important thing (which might be obvious from the context of this thread, but I still missed it) is that only the video was generated, not audio.

I was trying to find a song which contained something like "not getting rained on" and google AI just congratulated me that I avoided rain.

I started to use VSCode in 2026. Hardware caught up to it so it's not as slow as I remember (still slow, but fine). It's so noisy. Every cursor move highlights half the screen, at least 3 different searches active at the same time, it's a mess. Can't really write (or read) code in it. Using it for copilot.

> which failed to drive chrome so it wrote its own chrome driver protocol server

This is one of the things I hate the most. Super complicated workarounds which take loads of time (and sometimes money) for even the simplest problems. Human would pause and ask. I would blame harness, not the model though.

> This is at the top of Hacker News

Twice.

It obviously is? See above. Or do you not feel the clarification is important? AI being able to solve things no human bothered to try is great, but it is very different from AI solving things humans tried to solve and failed. And the latter is what pops to mind seeing these titles.

Or you could just take cash, as long as you don't need to drive.