phailhaus

7 ms·

91.
by phailhaus 1y ago
> I see no fundamental limitations How about the fact that AI is only trained to complete text and literally has no "mind" within which to conceive or reason about concepts? Fundamentally, it is only trained to sound like a h
92.
by phailhaus 1y ago
Yes, by several orders of magnitude. You can wipe out entire ecosystems if you mess around with insects, leading to mass famine and starvation. The "poison" we currently use is at least tested and regulated.
93.
by phailhaus 1y ago
Yes. Nobody has figured out how to get people to pay for social networks though, at scale. The free ones destroy the competition.
94.
by phailhaus 1y ago
Like I said: any system that optimizes for engagement has this problem. Advertising revenue scales with engagement, so engagement becomes optimized. Advertisers are not picking and choosing particular policy positions to place ads on. They&
95.
by phailhaus 1y ago
That has nothing to do with ads, that has to do with monetization. Every site needs to be monetized somehow. Ads scale with engagement, so engagement becomes optimized. Any monetization scheme that scales with engagement will have this issu
96.
by phailhaus 1y ago
This doesn't make sense, since it's advertisers who are the ones putting pressure on sites like Twitter to stop spreading extremist content. The problem is that humans are extremely willing to enter echo chambers where they are to
97.
by phailhaus 1y ago
We literally don't know. We don't understand how the brain stores concepts. It's not necessarily language: there are people that do not have an internal monologue, and yet they are still capable of higher level thinking.
98.
by phailhaus 1y ago
Feels like, but isn't. When you are reasoning things out, there is a brain with state that is actively modeling the problem. AI does no such thing, it produces text and then uses that text to condition the next text. If it isn't w
99.
by phailhaus 1y ago
The parent comment is talking about storing app state in the DOM. That's conflating data with presentation, which is a bad idea.
100.
by phailhaus 1y ago
Conflating presentation and data seems like a classically Bad Idea. Changes to presentation break your business logic.
101.
by phailhaus 1y ago
I get around that by sourcing a separate file in my config that I don't make public. Those are my company-specific settings.
102.
by phailhaus 1y ago
Instead of developing, I'm code reviewing. Hard to get into a flow state when claude is the one flowing, not me.
103.
by phailhaus 1y ago
And if a human failed on this question, that's because they weren't paying attention and made the same pattern matching mistake. But we're not paying the LLM to pattern match, we're paying them to answer correctly. Human
104.
by phailhaus 1y ago
> That's because the LLM assumes you mean for the host to open the door. LLM's cannot "assume". There is no thinking involved. It sees that the prompt looks like the monty hall problem and it just goes full steam ahea
105.
by phailhaus 1y ago
HN is resistant because at the end of the day, these are LLMs. They cannot and do not think. They generate plausible responses. Try this in your favorite LLM: "Suppose you're on a game show trying to win a car. There are three doo
106.
by phailhaus 1y ago
My work steps are too big to sit around pasting my repo into a text box every time I have a task. This is why integrated IDEs are taking off.
107.
by phailhaus 1y ago
There's no way I'm going to go through my repo dependency tree and paste twenty files into grok one by one.
108.
by phailhaus 1y ago
It cannot be better because Cursor looks across files, whereas with grok you'd be giving it a single one. Grok won't have any context about the rest of your repo, which makes it only useful for toy examples.
109.
by phailhaus 1y ago
I don't get this because someone has to work with the AI to get the job done. Those are the entry-level roles! The manager who's swamped with work sure as hell isn't going to do it.
110.
by phailhaus 1y ago
> 5. Protocols: As a Typescript guy, this seems very cosy & familiar. And not very Pythonic. I'm not sure how to feel. Isn't it super pythonic? One of the first things you learn about Python is that "everything is duck
111.
by phailhaus 2y ago
At some point, the user has to take responsibility. If you ask Google to save your home address, you can't then turn around and go "Chrome is collecting my physical address!!"
112.
by phailhaus 2y ago
But Chrome is not "collecting" this information secretly. Users are voluntarily giving this information to associate with their Google account. Chrome "collects" your physical address because it gives you the option of
113.
by phailhaus 2y ago
Can you clarify? Does Chrome track you, or do websites track you using cookies stored in Chrome?
114.
by phailhaus 2y ago
OP can just not use a TURN server, and it won't work for double-NAT traversal. I did this in a side project to avoid the security risk.
115.
by phailhaus 2y ago
Is Reddit too difficult to use for you? For me it's an endless feed of almost force-fed content. I'd actually prefer pagination to keep myself from going on for too long.
116.
by phailhaus 2y ago
Reddit doesn't control subreddit mods, and some videos not having sound does not make the site "unusable". That's why it's still so successful.
117.
by phailhaus 2y ago
They got rid of downvoting, which caused everyone's feed to devolve into garbage really fast.
118.
by phailhaus 2y ago
That's because Digg's changes fundamentally changed how users interact with the site and receive content. From an article at the time: > Protesting at the removal of the upcoming news page, the default setting of "My News&
119.
by phailhaus 2y ago
Who cares about security. As long as the libs are crying, right?
120.
by phailhaus 2y ago
Yeah I wonder if his experience is mostly using JavaScript, which is absolutely impossible to maintain at scale. Most of my team comes from primarily backend-dev roles and they've all grown to love TypeScript over Python.

More ›