I ran into 5 and 7 in a Flink app recently - was parsing a timestamp as a number first and then falling back to iso8601 string, which is what it was. The flamegraph showed 10% for the exception handling bit. While fixing that, also found repeated creation of datetimeformatter. Both were not in loops, but both were being done for every event, for 10s of 1000s of events every second.

(Perhaps a good library for timestamp code in data pipelines https://github.com/williame/TimeMillis)

Thanks! I'm using Instant.parse at present and this is supposedly 37x faster. Will definitely give it a try.

And report back please! :)

> documentation survives when it lives next to the code. 15+ years ago, this was pretty much the standard. Every decision - whether major or just a hack to handle a corner-case - used to be recorded in the code itself. Then tools like Jira and Confluence came in and these things moved to undiscoverable nooks and corners of the organization. AI search tools like Glean and Rovo have improved the discoverability, though I'd still prefer things to remain in the code.

Agree!

Rovo dev cli is pretty good though. Though that may just be because it talks to claude or openai in the backend.

I used it for a while a year or so ago when it was in beta and gave 20M free tokens daily

Rovo is backed by the typical LLM providers in general, Atlassian isn't training its own models.

They have enough enterprise customers to pay the bills for years though.

I loved their dry seasoning [1]. Bought some when I visited the bay area several years back and used to use it on everything from toasts to pasta. Sadly, haven't visited US since to be able to pick up some more :-(

1. https://www.amazon.com/Pepper-Plant-Seasoning-11-oz/dp/B01LY...

Reminds me of what bootstrap [1] was like around a decade ago. It's gotten quite a bit bloated since then though.

1. https://getbootstrap.com/

> A start-up I had been a part of in New York. Where I was the first employee. I had owned equity. They had eventually sold for $350 million.

What happened to the author's equity?

Obviously there is a ton unsaid in this blog post, but I just wanted to answer your question because it's exceedingly common for companies to be sold, sometimes for lots of money, and for common stock (which is what employees hold) to get wiped out. If the startup was sold for $350 million, but it received $350 or more million in funding, the investors get (some of) their money back, and employees get nothing. This happens all the time.

Again, I don't know what happened in the author's specific case, but think it's important to know that lots of startups have exits that can look big on paper but still are a wipeout for common equity.

Liquidation Preferences e.g.

Tried the app for Kannada. The "AI explanation" didn't work for any phrase, but the first lesson seemed to be quite useful otherwise.

Not just that, but potentially 17 other guilty caregivers have been cleared of suspicion based on the findings in that paper.

Yes he did. I attended a talk from him on the same, so that's how I know.