I've tried a variety of tools over the years and always end up going back to a low tech approach: spreadsheets. I set up my credit cards to email me after every transaction, which I then manually copy over to my spreadsheet with additional category information. Pivot tables give me a breakdown by category / month.

The added friction of having to manually copy everything over is a nice incentive to buy less stuff.

yeah big +1 to this. People are obsessed with automation, but doing a bit of manual work for stuff like budgeting and project management actually really helps cement it mentally I think (a bit like hand writing lecture notes, etc).

Spreadsheet + Pivot table is all you need in many cases.

If you're interviewing for staff level positions and above, you'll be discussing long term planning and some pretty tactical stuff where an NDA makes perfect sense. Might be a bit overkill for senior and below but meh.

Yeah, I could see it for any sort of high level position. I wasn't thinking in that context. I was thinking for just regular devs.

I can't help but feel like I lucked into easy mode given the general consensus that everything sucks right now. Applied to 10 places (mix of FAANG and small startups) after taking a year off from work. Heard back from 5, somehow got offers from all 5 and am starting at one of them next month.

It took a fair bit of prep to get back into things - especially when it comes to system design interviews - but otherwise everything went fine. Really enjoyed how some companies are trying non-leetcode approaches lately like code review and debugging sessions.

Which geographical market? Were you nervous after the time off?

I'm reading people who have been continually employed taking forever to find jobs.

I’m in Pittsburgh but didn't end up applying anywhere local. I was looking for remote-friendly positions to keep things flexible due to an upcoming move, which definitely limited my options. Out of the 50 or so companies I looked into, barely a third turned out to be remote friendly, with some only pretending to be (e.g. Meta seems to have gone out of their way to make remote suck for everyone involved by limiting it to staff+ on top of manager discretion).

I was definitely nervous about the time off being an issue. The usual stream of recruiter emails had pretty much dried up by the time I started looking for work, and I was fully expecting the search to take forever. It wasn’t an issue at all in the end - no one brought up the gap during interviews. I mentioned it in passing a few times when talking about why I left my previous job and mostly just got “dang that sounds like fun” from people. That being said, I have no idea if I would have gotten more responses to my initial inquiries without the gap on my resume.

That's excellent. And congrats for finding what you were looking for, and making some of us jealous about the remote part. You probably have a pretty solid skillset as well.

Now one third of them being remote-friendly isn't the worst stat in the world to hear. I was worried it would be far lower (at least in the Northeast). Was it much higher before?

Do you have experience interviewing / hiring? I wonder if people who find it easier to get offers just have a better idea of what the people on the other side of the table are looking for.

Another vote for Pocket Casts. The only thing I dislike about it is that the search feature can be janky at times, especially when it comes to non-English languages. No complaints other than that.

Because it's completely irrelevant.

and deceptive if not inaccurate. Meta's Model Cards specifically call out that they were trained on publicly available datasets and NOT any Meta user data.

For example: https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md

This model card seems at odds with what they claim you can and cannot opt out of [1]

> We use information that is publicly available online and licensed information. We also use information shared on Meta’s Products and services. This information could be things like posts or photos and their captions.

[1] https://www.facebook.com/privacy/genai/?_rdr

I suppose this is a bit of a special case but the US and Canadian governments both require you to provide the full list of trips from the last ten years when applying for citizenship and permanent residency. Timeline is incredibly helpful there.

Trips to other countries than where you lived, to be clearer for readers.

The goal is to know if you visited <insert questionable places>. They almost certainly already know, so it is also about concealment.

> that would necessitate constant retraining

You wouldn't necessarily need to retrain that frequently. If your model outputs hashes / vectors that can be used for searching, you just need to run inference on your new data as it comes in.

Definitely this, embeddings would be the modern approach.

The "modern" approach sounds like it would be a lot worse than this approach both in terms of input and runtime performance.

Trendy ("modern") is not necessarily better.

Judging from data on levels.fyi, definitely not the latter.

Amazing, relaxing, sound noise... ̿̿ ̿̿ ̿̿ ̿'̿'\̵͇̿̿\з=(•_•)=ε/̵͇̿̿/'̿'̿ ̿ ̿̿

Yes.

How did it go?

Yes.

How did it go?