I know very little about lawyering, but I could imagine a UW-alum or Seattle-area lawyer advising pro bono bc of generosity or good publicity on a very newsworthy case

Anyone on here friends with a UW-alum or Seattle-area lawyer who might be interested but doesn’t read HN?

Do American university students belong to unions?

It's very common in the UK. The most visible part of the unions is running social activities, often bars and events, but they can also provide legal advice to their members.

Not undergraduates. The "student union" in this context is 100% a social entity.

Graduate students sometimes are part of unions, but usually only if they're also employed by the university (somebody paying full tuition for an MBA probably isn't in a union, but a doctoral student teaching or doing research might be).

Undergrads doing part-time work at the university to pay bills (dining hall, bookstore, etc) could be, in theory, but probably aren't.

Graduate students (who teach, do research, and some administrative tasks for the university) occasionally do. What American schools usually have is a “student government” which is approximately 90% roleplaying as elected officials, and the remaining 10% is deciding which banquets they should host themselves to spend the small budget the university gives them.

Even if this student isn't a member, the local graduate student union (https://www.uaw4121.org/) would probably be a useful ally. All TAs are in the bargaining unit, and UW CSE has _a lot_ of undergrad TAs, so I wouldn't even be surprised if this student is a present (or former) member.

Possibly this is not the actual dress code? Or I'm missing something.

3.a. The following is acceptable for men players, captains, head of delegation. -- Suits, ties, dressy pants, trousers, jeans...

3.b. The following is NOT acceptable for men players, captains, head of delegation. -- Beach-wear slips, profanity and nude or semi-nude pictures printed on shirts, torn pants or jeans...

It's interesting that they are trying to enforce a different dress code for men and women. Surprising that's not seen some heat.

Thank you (and teMPOral) for these comments, this sounds potentially useful to me.

I hate to ask this, but I'm struggling to find any thorough posts or articles or papers about this, do you have any links you could point me toward?

Here is a short example that came up for me last week.

I had a set of documents I wanted to classify according a taxonomy that is well known (so it is exists in the training data of all the major llm models I tested)

If I have prompt like, `You are an expert classification system. Using the Classification Approach Foo, consider the following and output the category in JSON format, such as {"class":"bar"} `

This works ok, but it works much better if I tell it to output {"class":"bar", "reason": "baz"} and improved with some other approaches like adding "related_class" or "parent_category" which would otherwise be redundant.

Also including some few-shot examples helped, but the biggest benefit came from the "reason" field. Trying justification or other synonyms seems to produce the same output.

I suspect this is something similar to CoT.

Have you tested moving the "reason" field before the "class" field? That may encourage better CoT instead of having the model justify the class after it already picked it. Anecdotally, I saw a 5% boost in performance from a NER system from having the model output the entity's class at the end rather than the beginning.

This has worked for me for sentiment classification, i had a score and a summary field and the results were much more accurate.

Does it still work if the class field is output before the reason?

Yes. But, reason first did work (very slightly) better.

Great, thank you (and hedgehog), that makes sense

Speaking only for myself these ideas are a combination of things I've seen scanning new papers and informal discussions with other people working in the area. Feel free to shoot me an e-mail though, maybe I can point you somewhere more specific.

Edit: The "verbosity sink" name is inspired by the idea from the paper below although they're not actually at all the same thing.

https://arxiv.org/abs/2309.17453

I was in the S12 batch. A couple years later we were demonstrably failing. I emailed the general YC advice line or something similar with a question, and pg responded with a thoughtful comment that was genuinely useful

I don’t think there’s any chance he saw any economic value in us, he took that time because he cared about startups and well-meaning failures like us

It’s a small thing, yes, but I’ve always thought it said a lot about him

Agreed. In a similar vein, when our YC company was failing I had a great call with Jessica Livingston, and looking back it was purely moral support and not anything economically motivated.

Fwiw this feels deeply relevant to my usage of LLMs to structure data. I’d like exactly a good indicator of uncertainty for each bit of data.

I built a website to help people find summer camps for their kids.

https://campwing.camp/

Right now it’s just what i got up in a hurry last year, covers Seattle area only. Next year will cover SF and LA at least (and have a lot more features).

The data is quite granular (each row is one batch, say x age group at y location with z theme) and I do a lot of structuring data with prompts and now OpenAI’s Structured Outputs. I think the data gathering would have made this cost prohibitive otherwise.

I’m not engineer by training or work experience, so this also would have been impossible for me to build without ChatGPT’s help coding.

The notable part, to me:

“The memo cites a recent study, published in the journal of the Institute of Traffic Engineers (ITE), looking at Washington, D.C.’s implementation of no-turn-on-red restrictions at 100 intersections in 2019. Driver-to-driver conflicts were reduced by 97%, and vehicle-pedestrian conflicts were reduced by 92% following the installation of consistent no-turn-on-red (NTOR) signage.”

The Jesuit high school I went to had a class “Jesus of History, Christ of Faith”, that looked at what we do and don’t know about a historical Jesus (at least as of ~2000).

I’m sure I’d had my doubts before, but that class convinced me to stop believing for good.

Gotta hand it to the Jesuits for teaching that class though; all these years later and I still have a lot of respect for that order

I believe flybrand had to look up the word and then just wanted to save time for other folks (like me) who might also have needed to do so :)

I interpreted the capitalization of 'THEOLOGY' as an alarmed objection to the comparison.

The “learn to pronounce” made it seem pretty clear this was just copypasta, without any effort put into editing. The caps presumably just came from the dictionary’s styling. In this case indicating a more specific meaning in a particular field (theology).

Like glaugh, I appreciated the definition of a new-to-me word appearing inline :)

Yes, I was just trying to pass on something helpful based on my own ignorance.

Understandable. Actually if you Google the word theology is in all caps in the definition so the intention of the comment is unclear.

Yes, that was exactly it. I didn’t know the word and figured others might not.

Clavinova digital piano! I hated piano as a kid, largely bc of the lesson style (and plan on trying Suzuki method with my kids).

I’ve played guitar (poorly) for years and I’m shocked with how (relatively) easy it is on piano to do fun improvisation or learn songs I love (Radiohead’s Videotape to start, now trying Piano Man)

And digital means I can turn down the volume when kids are sleeping, use as a midi controller, etc. I’m not audiophilic enough to really care about the tonal difference between it and a real piano, I really can’t tell

Try out Pianoteq! Connect your digital piano to a computer (even Raspberry Pi) and try it out: https://www.modartt.com/pianoteq

This software will make your piano sound like a $20k grand piano.

ps - another fun project: adding an LED strip that responds to the piano keys - to create a beautiful visual rainbow when you play.

There are free vsts you can use instead of pianoteq as well.

On linux I set up my midi controller with Carla to use Salamander Piano [0]:

[0]: https://sfzinstruments.github.io/pianos/salamander

Yes, but it starts at $149. Unless you're doing music professionally, that is rather steep when free ones like Salamander exist.

Does it have weighted keys? Do you have a link to the model you purchased?