jamesrom

How would you differentiate between real, valid criticism and 'altright cope'?

State of the art is a bag of beans:

https://dys2p.com/en/2021-12-tamper-evident-protection.html

What is especially egregious about the title is that "blending" requires stuff to blend. It's not all you need! It makes no sense at all.

The best way to share a dollar with everyone is to burn it.

http://www.aaronsw.com/weblog/tdk

However, it might be a tad hypocritical to name your company "OpenAI" in such arrangements.

The contents of a $30 book is available here, free.

Yet the comment section will endure much complaining by those who feel entitled that the layout, formatting, and typesetting be given away for free, too.

Anything that helps Terence Tao find bugs in his papers is hard and reliable in my books.

Is anyone else distracted by all the emphasis, italics, underlines, and highlighting?

When everything stands out, nothing does.

Sorry for the off-topic comment, but I do think this is valuable feedback for the author.

You can report SMS spam from the Messages app.

https://support.apple.com/en-gb/guide/iphone/iph203ab0be4/io...

This could be confirmation bias. All the times the progress bar just quietly did it's job, gave you some indication of progress and loaded without issue you are not likely to remember.

But the moment it's "wrong" it's annoying.

Why not use a functional programming language if you need FP?

> and need to be fairly large to be efficient

This statement is not true. Actually the inverse is true, the smaller you can make the disks, the better the efficiency.

See https://onlinelibrary.wiley.com/doi/full/10.1002/ese3.1417

> As disc outer diameter decreases from 100 to 50 and 20 mm, the isentropic efficiency of the miniaturized turbines increases significantly from 21.46% to 30.78% and 32.21%, respectively.

A use-case for -0.

https://en.wikipedia.org/wiki/Negative_zero

Oh yeah! How could I forget! Added.

Hyper.app: https://hyper.is/

Hyper.sh: https://github.com/hyperhq

hyperapp: https://github.com/hyperapp/hyperapp

HyperApp: https://itunes.apple.com/app/apple-store/id1179750280?mt=8

HyperTerminal: https://www.hilgraeve.com/hyperterminal/

HyperShell: https://github.com/holepunchto/hypershell

Adding this to the list https://news.ycombinator.com/item?id=14977382

You said a lot of stuff, but can you explain, specifically, one thing you want your TV to do that a simple PC display + GPU could not do?

All docs are generated. Now some are generated by silicon.

Code is the ultimate reference for understanding a project, while documentation is often neglected, outdated, or incorrect. It can also be difficult to keep up to date.

An LLM may not fully comprehend the code like the original author, but it can offer a different perspective that may be valuable. The only argument I've seen against LLMs is that it may encourage laziness, but this is a flawed argument similar to those made against the printing press, which was said to make people illiterate.

As a reader of the docs: does it require discipline to refer back to the code when needed. Yes, but this is no different from the discipline required to write documentation in the first place. But with a key difference, the discipline is shifted from author to reader.

Doesn't sound like a compelling reason against UTC. Add a leap second every 50 years for relativity.

So the moon gets a leap second every 50 years. Why does that prevent you from using UTC?

> For completeness I include gravity because it's the only other fundamental interaction we arguably directly sense (inner ear fluid and fine hairs are part of the same sensory organ, unlike our skin and the sun).

Not sure why the organ doing the sensing matters. I'm talking about thermoreceptors in the skin, obviously.

But just to be clear, I'm saying that you're actually sensing electromagnetic force in your inner ear _caused by_ gravity. Not gravity itself.

Even if the hairs bend due to gravity alone, that force upon the sensory receptors is an electromagnetic one.

But if you consider that 'direct' then you must also consider sensing of the weak force, as 'direct'. Whether it be via thermoreceptors or your retina.

tldr

gravity : bending hairs :: weak interaction : radium glow

What do you mean by directly? Gravity is only sensed via an interaction between fluids and fine hairs in your inner ear.

By that definition, we can also detect the weak interaction by feeling the sun on our face or the warmth of radioactive material.

I'm sorry, but stop it with the "thou shalt not use black on white". _That's_ the predictable evergreen amateur design bro take.

Once you realize that "appeal to nature" is a fallacy, you will be free.

Agreed, there are many reasons to choose pure black and pure white.

Everything about modern design is "unnatural", this design rule is the "appeal to nature" fallacy manifest.

If you need proof that beautiful things can be black and white: https://gwern.net/

(also, OP blog is _so close_ to pure black and pure white, it's almost imperceptible)

Turns out, we only needed to wait a couple of days: https://neeva.com/blog/introducing-neevaai

Maybe. But that doesn't mean it wont be a lot better than current-gen search by then.

Do you have any reason to think LLM cannot reference published work?

Let's check back on this in a few years. The idea that a search engine can cite experts but a LLM cannot sounds wrong to me.

I'd love for you to explain how a search engine does have any idea of truth.

The main argument of this article is:

1. ChatGPT gets tripped up in some edge conditions.

2. Search engines provide some interesting features that ChatGPT doesn't.

These are both bad assumptions looking forward:

1. It's evident that ChatGPT or other chat-based text generation will only improve over time. These kinds of quirky edge cases will _always_ exist, but will be harder and harder to contrive.

2. These sorts of features, categorization, images, etc can be discovered via conversation. And there's no reason to think these features are exclusive to a search product. Code snippets embeds are already built into ChatGPT, it's not hard to imagine other useful embeds.

Finally, I'd ask the author to think of it this way:

As more text is generated, and those pages end up at the top of search results, the quality of the search results approaches that of a text generation model anyway. Search will soon be the unnecessary middleman.