jacobsimon

Yeah maybe it's a grass-is-greener situation, but from a quick look at SwiftUI and Jetpack Compose, they both seem to have a lot more in common with React than HTML/CSS itself. They seem to encourage functional UI that combines state management, event handling, and view layout. They have primitives like `Text`, `Row`, `Column`, `Spacer`, etc.

I do think the web is a lot easier to learn and the on-ramp is so much faster compared to older mobile UI frameworks, so it makes sense that they're sort of evolving toward the same patterns.

Not OP but at the end of the day, custom components like this still have to be rendered down to primitive HTML and CSS, and are plagued by all sorts of browser-specific issues and conventions that make higher level abstraction way more difficult than they need to be.

As a random example, HTML specifies that clicking a `<button>` inside a `<form>` automatically submits a POST request to the current URL. We can't change that because it will break a lot of webpages that rely on this default behavior, so now every component system built on top of <form> or <button> needs to prevent the default browser behavior, and there are so many more examples of things like this.

There are certainly lots of great component libraries out there for various frameworks that help with this, but I agree with the parent comment that it's still not the kind of progress we should have had by this point. I think it's unfortunate that we have to stick to HTML/CSS as the foundation for everything to maintain compatibility and accessibility, but that's how the web works I suppose. If we were to design the web from the ground up today based on what people use it for, it would look pretty different I think.

I like React but I actually agree with you re: JSX, it helps me to remember that JSX is 'just javascript' - it's a syntactical shorthand for creating a component tree, which then happens to be rendered into HTML.

I had the same thought this morning and was debating selling my nvda stock when I saw this - feels like they are well-positioned right now, as with crypto a few years ago, but if there were an efficiency breakthrough that allowed commodity CPUs to do the inference instead, this advantage could vanish quickly.

In other words - it doesn’t have to be someone’s genius plan, it could have just been an unintelligent mistake

Not OP but I think both perspectives are valid. Real estate agents are not incentivized to increase your house price at the margin as much as the owner, they would rather just keep their fee high and sell your house at an ok price.

Maybe real estate agents should structure their commission more like M&A bankers and account executives, which often have tiered or ramped fees that encourage the best prices, e.g. if the selling price is below X you get 5%, everything above X you get 10%, etc

I think the other key question is causality - do people who get less deep sleep develop these diseases, or do people who are developing these diseases get less deep sleep, or both?

I wonder if they could design an interventional study in which people who are predisposed to dementia are given more sleep or melatonin, and then demonstrate whether it protects against the disease.

My guess is that the firing rate of biological neurons more or less simplifies to the activation in an artificial neuron. Higher firing rate = higher activation.

Plus: https://x.com/ylecun/status/1583872918634655744?s=20

I don't know why I'm still responding to this thread 24 hours later, but just thought I'd add this tweet from Le Cun: "Neuroscience greatly influenced me (there is a direct line from Hubel & Wiesel to ConvNets) and Geoff Hinton. And the whole idea of neural nets and learning by adjusting synaptic weights clearly comes from neuroscience."

https://x.com/ylecun/status/1583872918634655744?s=20

TIL

Yep! Trees, tree structures, tree diagrams have been regularly in use since the 1700s as a way of defining relationships. https://en.wikipedia.org/wiki/Tree_structure

There’s also a pretty large link between the formal representation of language using syntax trees, which was being formalized by linguists and by programming language developers around the same time: https://en.wikipedia.org/wiki/Formal_language?wprov=sfti1

I mean, I know you’re just providing an analogy, but people are still studying the physics of bird flight and we’re nowhere close to building machines yet that can maneuver the way birds can. https://www.quantamagazine.org/geometric-analysis-reveals-ho...

Surely you are trolling me now. There is a very clear biological inspiration mentioned in this paper: they literally define a CNN as having “receptive fields” and then they cite the same Hubel & Wiesel research mentioned before multiple times. LeCun mentions their research in papers even earlier in the 80s as well, during which they were awarded the Nobel prize for their research on the visual system. Of course there is also a lot of computational and mathematical research that was ongoing simultaneously, but to say that there is “no inspiration whatsoever” is pretty far from the truth.

This is a really weird take. There is such a long history of shared insights between biology and neural network research, and to say they’re unrelated or can’t take inspiration from one another is bizarre.

> The concept of CNNs didn't come from biology

I just opened a survey paper on CNNs and literally the first sentence of the paper reads:

> “Convolutional Neural Network (CNN) is a well-known deep learning architecture inspired by the natural visual perception mechanism of the living creatures. In 1959, Hubel & Wiesel [1] found that cells in animal visual cortex are responsible for detecting light in receptive fields. Inspired by this discovery…”

Source: https://arxiv.org/pdf/1512.07108.pdf%C3%A3%E2%82%AC%E2%80%9A

Oops I got curious and went down a big rabbit hole.

There’s a lot of interesting observational research on this topic, and in particular I found this paper[1] interesting because it shows a similar link with Tylenol and autism in young infants, but did not find the same effect for ibuprofen, which to me almost rules out fever/infection as the root cause.

There was another paper I saw that even suggested the anti-vaxxer conspiracies may stem from the fact that doctors and parents give Tylenol to their kids to manage the side effect of some vaccines.

Interestingly, there is also a separate body of research[2] that has shown experimentally that acetaminophen causes emotional blunting and reduced empathy in adults - could be related?

1. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5044872/

2. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6455058/

Maybe Apple should add a similar distinction between transactional and marketing notifications as we have with email.

Not an expert but the wiki article says he proposed a link with acetaminophen, which has recently been demonstrated by papers and spawned a lawsuit: https://www.spectrumnews.org/news/acetaminophen-on-trial-ove...

Let’s say I want to know the conversion rate of my payment page. I need to know how many unique visitors viewed a page, and how many of them went on to complete a payment. I’m pretty sure this is not possible without a unique identifier. And if you are using (ip address, user agent) as a proxy for such an identifier, is that any better or legal than using a cookie in the first place?

Can you explain how server side analytics works without cookies?

How would that work with video editing? Like if someone records something and then trims it for length or needs to combine multiple streams. Seems like hardware level verification only goes so far.

Super interesting - this almost sounds like evidence of linguistic determinism / relativism, i.e. the idea that our language influences how we perceive and think. Is that what you’re thinking too?

This is a really good observation. If I can add on, I feel like a lot of the disagreements in modern front end development boil down to this: Some people actually like developing with HTML/CSS, and other people see it as an implementation detail and would rather abstract it away with JavaScript.

People in the first camp tend to like Vue, CSS/Sass, Htmx, server rendering, etc.

People in the second camp like React and React Native, CSS-in-JS, Tailwind, etc.

I just came here to say that I actually find phenylephrine nasal spray quite effective - but I use it for a different use case: stopping nosebleeds quickly. It's quite effective at that and came recommended by an ENT, so I do believe it's an effective vasoconstrictor when applied locally and I'm glad it's available.

recommend tailoring your resume to each job and highlighting the experiences/projects that are most relevant for that role :)

also applying to newly posted positions and using referrals when you can

Been wanting something like this for years over at Exponent. We do a lot of targeted and dynamic emails to our users based on their role and progress through our courses, and it’s a pain to not only synchronize the database with the mail provider, but to agree on a system for capturing the state of the user that works with the platform - we’ve ended up creating a complex system of tags that has to be maintained across both systems.

Re: non technical users, one idea would be an ability to save “audiences” using smaller SQL queries, that way non-technical collaborators could easily send a campaign to a well-defined group of users, and update the group definition once instead of across all your emails.

If you could infer the structure of the connected database automatically and control which tables are visible in the product, people might find that really helpful too (check out Metabase as an example product that does a great job at querying data sources with UI and raw SQL)

I don’t think they changed anything drastic. I’m still able to do the terminal simulation prompt and pretty much anything else from before. I think they improved the guardrails around the responses so you have to try a bit harder, but I wouldn’t say it’s worse.

Great idea, I was going to add basic retries but didn’t think to include the error.

Any other features you’d expect in a prompt builder like this? I’m tempted to add lots of other utility methods like classify(), summarize(), language(), etc

Coincidentally, I just published this JS library[1] over the weekend that helps prompt LLMs to return typed JSON data and validates it for you. Would love feedback on it if this is something people here are interested in. Haven’t played around with the new API yet but I think this is super exciting stuff!

[1] https://github.com/jacobsimon/prompting