elijahbenizzy

While simultaneously achieving record revenue/profit?

That looks slick! Didn’t have any idea shiny was in python, but IIRC data scientists I’ve worked with in the past love it.

Yep agreed, I really found it good for the first something — e.g. a more advanced figma. Past that it was a bit of a gamble.

On the same team — I’ve really enjoyed strsmlkt and built an app (that made it to the HN front page!) https://share.streamlit.io/-/auth/app?redirect_uri=https%3A%....

That said, I’ve found it to be clunky in a few ways:

1. State management is not intuitive — anything that involves rendering complex state/reacting at it changes requires a ton of session state type stuff.

2. I’ve very often had to go into css/markdown hacks

3. The control flow of when it’s running/executing makes for clunky interactions

4. Some API decisions are painful. For instance, a block str comment in the file will render as st.write

5. Long running (or even not instant tasks) make for some annoying engineering challenges — the UI is tied to the compute, so you have to be smart about juggling it

6. Caching is not smart — it’s quite useful, but it’s easy to get lost in caching/state

I’ve been dreaming about streamlit but a little more state/function oriented. E.g. react-based. Not far from fastUI

These are both great tools for prototyping, but I’ve found that I’ll pull up tailwind + some UI framework when I want something slick and it’ll take a few times as long.

There's an argument that the AI tech bubble is just a continuation from the value-add of automating everything and the productivity/GDP-increase that causes. It is just that investors are so skittish that they'll only let the floodgates open if its "sexy", so AI comes in and restarts the money machine.

There's an interesting mental model I've been toying with. At what point do LLMs just become circuit-shaped NNs with stochastic gradient descent backing them?

E.G. are we just determining the best program by rearranging 1s and 0s?

You've built observable but for python. Love it!

This is true AI! Nice work!

Interesting points about the nuances of a $450 bag that’s rather impractical when it’s more a story than a justification. I’m just here to comment that this was a lovely read, thanks for posting!

Interesting! I had never heard about that.

I think that if you look at a history of social changes you’ll very likely find that they all started with successful persuasion campaigns. Like it or not, if you want to preserve the rights we hold dear (as you mentioned above), you’re going to have to find people that agree with you, hopefully through facts and reason. While it’s possible to preserve rights solely through a well-built technocratic/representative government, the will of the people will always bubble up in some way, and you owe it to your causes to do everything you can to change that will/find people that agree. Whether this applies to fireworks or not, left as an exercise to the reader.

I disagree with large parts of what you said, but if my and others’ anger towards fireworks (which I actually do quite enjoy) results in a small tax on firework sales to fund veteran causes, in which I am a staunch believer, then not gonna lie I’d be thrilled.

See comments above -- its not logic for banning, its a coalition-building tactic, which is generally employed in most decisions we make/persuasions we carry out.

Banning things, adding taxes for things, providing incentives are all tools we have to manage externalities. We generally have a budget for how many restrictions we can impose, and we've banned plenty of things (such as slavery). Obviously we don't just go on a banning spree, but the argument that "banning is bad, America doesn't do it" is, IMO, not realistic or how our country has always operated.

This is tactical in nature. In a democracy, you need to find a coalition to support what you want (either directly to vote for it, or indirectly to pressure their representatives).

All decisions of this sort come from a passionate group of people. I'm suggesting dog owners. This does not extend to how to make decisions.

Oh don't get me started on outdoor cats...

This is why we also have the arguments "Think of the dogs!" and "Think of the veterans!", which no empathetic person can be against (even though, yes, these arguments employ an array of manipulation tactics...).

Sure, but one does not need to convince the fans of fireworks, rather one needs to convince enough fans of dogs (of which there are quite a few)

Strong disagree, but there is clearly an argument for everyone here :)

I honestly believe we should ban fireworks and firework shows across the country. Not only are they (apparently) bad for the environment and likely to cause wildfires*, have the potential to trigger PTSD**, but they also make the large majority of dogs terrified for at least two nights of the year. As humans, our symbiotic relationship with dogs is one of our greatest assets -- we owe them better than this. Drone shows + music are just as cool.

* https://www.readyforwildfire.org/more/fireworks-safety/

* https://www.kron4.com/news/bay-area/fireworks-spark-dozens-o...

* https://www.nfpa.org/education-and-research/home-fire-safety...

** https://www.ptsd.va.gov/understand/what/fireworks_ptsd.asp

** https://www.va.gov/hines-health-care/stories/how-your-firewo...

This has the same feeling as leaking AI-generated text back to the training data...

Wonderful! Yeah, admittedly, I haven't sat out and drawn it out (my process for learning algorithms is often just do it by hand until it intuitively makes sense), but it did strike me as pretty straightforward on first glance. Thanks!

I know LLMs are all the rage, but I’m genuinely shocked at the new fastest path algo for negative weight graphs. That’s awesome!

Yeah! Kind of terrifying... The only thing giving me comfort here is that these were all done in a single day, and I can't help but wonder if there's some bug or correlation. That said, even if its a bug in the model that they rolled back, it's still scary.

Wasn’t interpreted that way! And yeah, part of it is stochastic, but part is just that it has crazy associations and no social cues so it kinda just does whatever.

Good point -- in this case I'm using the URL it passes back, which was the only information it had. The original one was passed either as base64 encoded or a URL. IIRC the URL they passed back was random-feeling, but I'd have to look to check it out. Its possible that the initial URL encoded a little more, depending on how the filename works, but my guess is that the model is good enough that, say, "royce_hall.jpeg" wouldn't change it too drastically, especially not over time...

Yeah, its kind of crazy sometimes. Quite possibly training data on DeviantArt?

Hey now, that's a latte I made :) But yes, agreed, B- art without question. The caption is based on the original image, which is one I snapped a few mornings ago. Feels like it only moves towards an overdramatic representation, which probably has to do with its training data? Also likely with how I prompted it -- you can see the prompt + caption + embeddings here: https://d1lf8m1wnxcl0a.cloudfront.net/latte_2_20231217/metad... (it'll download a json).

Note the prompt is meant to keep it interesting (ish) -- this is all parameterizable by the code, but my guess is if I removed the second sentence it would be a little more boring:

"Please provide a caption for this image. The caption should be obsessively descriptive"

Looks to be back up and running faster! Possibly cause this is no longer on the front page. Also I have confirmation that someone at snowflake is looking into it :) Enjoy the dataset -- still uploading...

Unfortunately nope, see the description above. Costs about $15/sequence of 150, but if you have one you'd like to see I'm happy to kick it off...

That said, all the instructions for getting started and running your own are there (including a jupyter notebook!). If you have some data you generate I'll happily add it to the platform.