adlpz

Any context on this? I see lots of recent works. Is this just an extension of the sci-hub concept to just everything?

Anyone has tried self hosting the server but sending through a third party like Amazon SES?

This hits home quite clearly.

I'm just starting to think that coding won't ever be a fun thing anymore. Just another boring, mind-numbing thing I do for money.

I'll get a hobby I guess.

It also frustrates me much more.

A hard bug used to be frustrating, but also a bit challenging.

Not anymore, really.

I'm literally doing the same.

Feels wrong but it's the closest I got.

100%. It's truly infuriating, but going back to my normal workflow now feels somehow wasteful. Like I'm able to rely on a team of devs for busywork and just deciding to do it by hand and overcharge my clients (in time).

I've had a few instances where I've been able to go into flow state with 3 concurrent sessions on fully disjoint projects. Feels like context switching is easier when there's almost zero cognitive overlap between the projects in terms of actual features, but the stacks are all similar.

But for sure I haven't even come closer to these glorious flow sessions fully locked into coding some clever bit of logic.

So, here for the suggestions too.

Interesting article. I've also noticed the effect, just never seen it expressed and contextualized.

Not that it really needs any historical context and narrative flair to be useful:

Leave in-progress things at an interesting/engaging point so it's easier to continue later.

And the article pointed me to https://www.mirjamdonath.com/the-hemingway-challenge/ which lists 50 books-to-read (or 49 really) based on a bunch of Hemingway's quotes.

May do the challenge myself.

It's a bit like general web browsing.

The internet is full of pure nonsense, quack theories and deliberate fake news.

Humans created those.

The LLMs essentially regurgitate that, and on top they hallucinate the most random stuff.

But in essence the sort of information hygiene practices needed are the same.

I guess the issue is the deliver method. Conversation is intrinsically felt as more "trustworthy".

Also, AI is for all intents and purposes already indistinguishable from magic. So in that context is hard for non-technical people to keep their guard up.

Ah yes, unnecessarily skeuomorphic technobabble is the best kind of technobabble.

That's called voting with your wallet.

That would be amazing!

Very cool work. This hits the nail on what I've always kind of needed. Or, really, what I've kludged from an assortment of ever changing tools and glues.

My issue: what about the other tasks? With this tool I'd have my task list and my email-task list, in two separate places. Not ideal.

I know it's probably way out of your scope to also build a good general purpose task manager. But I feel like another kanban to manage is one kanban too many.

To be honest that's what I hoped Notion Mail would be. I've landed back to Notion as for task management and something that would autosync my emails so I can "see" them (as in, move them around in a kanban and add notes to) in Notion itself, essentially making them tasks would be the perfect system.

Anyway, I do love the people focus. Makes much more sense than any other label-based attempt.

Best of luck!

Go read a book and grow some tomatoes.

May as well be bored about steam turbines or refrigeration.

This posts reeks of a worrying lack of agency.

This is the funniest thing I've read in a while. Thanks.

I've run isopropyl alcohol in an ultrasonic cleaner.

It did get significantly hot after a while. And smelly, so it was surely aerosoling (is that a word?) somewhat.

I did it outdoors and away from anything that could catch fire because it felt rightly sketchy.

Looks cool! If I may, a couple of feature suggestions:

1. You mentioned it already but bookmarks. In particular bookmarks that can be added through iOS Shortcuts (so we can each build whatever crazy automation we need) and that can be exported (so we can keep LARPing productivity in our Obsidian vaults).

2. An "undo" function, somehow. I know this is weird but I've misclicked the seek bar so many times (esp. in the lock screen) and lost where I was... a solution for that would be so cool.

I've been using the same podcasts app for... well over 10 years. But yours looks really cool, I may just switch.

This is lovely. And it's great to have it both in English and Spanish because it makes it much easier to guess the sounds from the explanations as you can compare.

New life goal unlocked: live in a farm away from any computers and learn latin.

Just talking about my own usage, I follow lots of people in the tech space, indie hackers, AI researchers, etc. By checking Twitter from time to time I learn about new tools, libraries and models, check out cool stuff made by others, etc.

It's literally the same reason behind reading HN. A HN with a higher ground noise but much more real-time, interactive and orders of magnitude higher bandwidth.

You just need to curate things very hard so make sure you keep enough signal above the nose to be worth it.

Except it's not performance usually, it's just features. And then it gets bloated. And someone things they don't need all that crap.

But turns out they did.

It's time we accept we won't be much more than a glorified theme park in a few decades.

Either that, or we stop complaining and start acting.

Maybe it's that we're just the same as these EGAIR people?

Any background on how the code got leaked? Insider? Hacked servers?

It's a 1/3 scale prototype so I guess through some creative maths the speed of sound may as well be 1/3?

Supersonic flight isn't necessarily for the 0.1%, just the 1% (or whatever other made up number) whose price per hour times the travel time reductions is greater than the added cost of the fare.

I see where we differ in opinion. I do believe that the human mind is just a very particular material arrangement. Therefore while different in scale to an LLM and likely also differing in structure, I do think both systems are fundamentally in the same category.

Given that, we could never agree. But that's fine.

And by the way, I actually hope you're right.

It is meaningless to me because the term is imprecise. Is it precise when I do something bone headed and say "sorry I wasn't thinking"? Do cats think? Do worms?

Depends on what you mean by thinking in that particular case.

In my opinion what LLMs do is close enough to thinking for some definition of the word.

And, by the way, I think there's no need to be unpleasant to state your point.

My point precisely. Those are all vague terms. Saying that "neural nerworks do not think" is as meaningless as any equivalent (or opposite) statement on any other system including any number of neurons, a whole brain or a person.

It's all semantics.

Do neurons think? Do a bunch of neurons?

Is this semantics?

Any relation to OpenAI's rumored Q* (i.e. q-star) model? Authors of this paper don't seem affiliated.

Just a name coincidence?

As far as I've seen, for APIs that don't "consume" any resources, the usual schema is to charge a fixed fee.

Either that or it's just included part of a higher pricing tier. Of course, there can be a quota restricting how many API calls are allowed per period.

Charging per call feels weird when the cost of a call is essentially zero. You only really see that in things that are computationally expensive like LLMs, image optimization, etc.

MacOS