eurekin

Right there with the infamous $18 BMW heated seats subscription

Also context size significantly impacts ram/vram usage and in programming those chats get big quickly

Was very surprised it wasn't mentioned

They have? Is there a write up about that?

That's exactly what everybody advised me against doing - finetuning on own projects. Got really discouraged and stopped. So glad someone has done it!

That's perfect, thank you!

Guessing that a ultrasonic microphone from a bioaccoustics site should fit the bill - otherwise, what could researchers pick themselves?

This one, for example, picks up to 200kHz for €1050:

https://avisoft.com/ultrasound-microphones/cm24-cmpa/

Is the sound part real? What frequencies are used to communicate stress? Is this in range of anything I could connect to a raspberry pico or arduino? My flowers desperately need answers :D

Technically, wouldn't a simple "Hold on, I'll call you back" test call stop that?

I distinctly remember the first time a lecturer used the "dx/dt" "symbol" in normal algebraic operations (that is, multiply both equation sides by dt and so on). I was so shocked it's actually not a elaborate differentation symbol, but something with actual division. Next time it was similar with integration, where the dx was substituted by some other function of du.

I swear I treated those as some grammar token, which doesn't hold any real meaning. I've been using those as such for years before.

Corridor Channel had one great example: https://youtu.be/GaGcLhhhbDs?si=vDyeayLf8EAoE0gf&t=442

Above includes the explanation. Final result is here:

https://youtu.be/GaGcLhhhbDs?si=eoTniegWK-AVFoaF&t=751

If that technically worked, I could see that being a next unicorn.

Bisphenol A have a carbon ring structure that resembles hormones like the estrogen. It can bind to the estrogen receptors and potentially increase risk of hormone related cancers.

In the age of LLM's, I with there was some way to automatically pull videos about particular topic, extract the information out of the transcript (skip the clickbait as well) and deduplicate. I spend so much time watching videos just to conclude "ah right, it's actually based on the same info previous 3 videos"

Anybody has experience and is it actually useful?

Heard a lot of "it captures mostly face muscle contractions" and "capturing a brain signal with it is like listening to a whisper being spoken in a grand canyon, while being miles away"

I work around that (with limited success) by prepending the latest docs (migration guides, how-tos) into the conversation. gpt-4o picks that up perfectly

there already are mirror like beetles, maybe weeds will catch up to that :D

I'm waiting exactly for this, gpt4 trips up a lot with blender currently (nonsensical order of operations etc.)

Same... I especially love pasting in a screenshot along the outerHtml from chrome devtools and ask "make those aligned on the same line, the thing with the text on the right" and it finds the right combination of flex and margin values on elements.

Feels like a huge burden lifted off my shoulders

Of course one can dumb down the requirements so that it will handle, but what about "once basic auth is added, check which endpoints should require it and by what clients" - any real work is out of the scope currently

Ok, so you're a library developer and create a greenfield API.

What do you do, in order for chatgpt to be able to pick up your library and it's patterns? What obstacles I see in this scenario:

* Base models takes months and millions to train

* RLHF supposedly can add knowledge, but it's disputed to mostly "change style"

* What incentive will OpenAI have to include your particular library's documentation?

I imagine, if that library starts being really popular, a lot of other code will include examples how to you use it. What about before that?

Including new knowledge always lags (are there two gpt updates per year? maybe a up to 4, but not really significantly more) few months, so what about a fast moving agile greenfield project? It could cause frustration in LLM users (I know I have been bitten a lot by some python library changes already).

It seems that it's just another tool in the box for humans to use. In far far future maybe, when we somehow get around those millions of dollars for fine tuning (doubtful) and/or libraries simply stop changing.

But still, put any really not small code base into 120k token context and see how easy both gpt and cluade opus trip up on themselves. It's amazing, when it works, but currently it's a roll of a dice still

It's a great and very useful tool, but the hype around just feels significantly blown out of proportion.

I also built multiple things with it and I always came to a point, where it just couldn't handle anything slightly larger than a mvp, or a non guided change requiring editing multiple files at once

I'm using chatgpt daily, but cannot imagine it doing any real software development work for multiple reasons.

It's great as a non offensive stack overflow replacement, but just look at aider benchmarks (amazing work by the way): most capable models really struggle to make basic changes to a real code base.

Does anybody actually using it in practice believes the hype? I thought the hype is just another theatre for investors

EDIT: just some points taken from: https://aider.chat/docs/leaderboards/#code-editing-leaderboa...

- The metric "Percent completed correctly" maxes out at 72.9% with gpt4o, while at the same time, giving out correctly formatted output only 96.2% of the time.

- Benchmark suite is based on https://github.com/exercism/python, which very likely is a part of the training material already! In real code bases, no LLM would have the advantage of seeing new or proprietary code

Yeah, that's what I'm afraid

Thank you! Is the detection accurate enough, or simply the observation conclusions are not that sensitive to minor errors?

That makes me want to revisit my previous idea: boiling soup spillage detector. I once had a google meeting with a cooking soup to keep an eye on it and thought, heck, that seems like a nice exercise for a visual detector finetune

Plus, being a NN it might be possible to detect a foreign object with relative ease (comparing to the classic computer vision); like a rat

After years of wondering, I have to ask.

What are real life actual useful cases for this tech?

I can imagine in manufacturing: detecting defects or layout mismatch - that's one.

Is there any open source project that uses a image recognition library to achieve any useful task? All I've seen from board partners seem to at most provide very simple demos, where a box with label is drawn around an object. Who actually is using that information, how and for what?

I've also been a part of the Kinect craze and made 3 demos (games mostly) using their SDK and still have a very hard time defending this tech in eyes of coworkers that only see this as a surveillance tech

Multiyear long spotify listener.

I think it's already flooded with spam from non real artists. I had weekly discovery playlists, where I did downvote all of proposed songs... Each new set was still coming with most boring, uninspired, flat and predictable structure (and abstract cover art), which for me is an exact equivalent of those NFT images.

The images are designed to be a set of replaceable elements that have to follow the same "joint structure". Once you see it, all charm is lost and that vague "why does it look so funny" feeling is simply replaced with disappointment.

That reminds me of one case, where an author of software saw any improvement or genuine feedback as a personal attack - directed at him and his "legacy". Can't go into specifics, but only features directly or indirectly praising the original software were allowed. I was student/intern at the time and learned about that the hard way.

Still, one of most important teaching moment to this day.

What did the management specifically do? (500 user)