Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Training our own AI models

posthog.com

Post reply on HN
Print view View original

151–156 of 156 posts

Re: Training our own AI models

#151

Post by kiwicopple »

rafael-lua wrote: post #138
Waterluvian wrote: post #46

PostHog was a system we set up once, generally don't think about, and review from time to time, providing some occasional value. It was mostly harmless to leave around. But it's apparently yet one more thing we have to be actively suspicious of as it defaults towards an intolerable state. So it's easier to just rip it out of the system and move on.

PostHog was wonderful as an analytics solution, with its developer-first approach, good tools, and nice pricing. At this point, I lost count of how many times in the last two decades I fell for it, although I'm more used to it now. Companies that grow into success and change. With the AI frenzy, PostHog also started going all-in, and not only that, but they seem to be exploring no-coding tools and whatnot. Supabase i…

> Supabase is another one that was cool, but now it is in the AI abyss.

(supabase ceo) I'm curious what gives you this impression? it's true that a lot of platforms build on top of Supabase, but our focus is still just building the best Postgres platform we can (OrioleDB, Multigres, etc)

admittedly we've also had to shift a lot of our tooling to things like Claude/Codex, but that's more reflective of how people are building now

Re: Training our own AI models

#152

Post by rafael-lua »

kiwicopple wrote: post #151
rafael-lua wrote: post #138

Earlier quoted context omitted.

PostHog was wonderful as an analytics solution, with its developer-first approach, good tools, and nice pricing. At this point, I lost count of how many times in the last two decades I fell for it, although I'm more used to it now. Companies that grow into success and change. With the AI frenzy, PostHog also started going all-in, and not only that, but they seem to be exploring no-coding tools and whatnot. Supabase i…

> Supabase is another one that was cool, but now it is in the AI abyss. (supabase ceo) I'm curious what gives you this impression? it's true that a lot of platforms build on top of Supabase, but our focus is still just building the best Postgres platform we can (OrioleDB, Multigres, etc) admittedly we've also had to shift a lot of our tooling to things like Claude/Codex, but that's more reflective of how people are b…

[dead]

Re: Training our own AI models

#153

Post by micromacrofoot »

abustamam wrote: post #136
micromacrofoot wrote: post #132

Earlier quoted context omitted.

has it not always been this way?

Two things can be true. A shitty thing can prevail for a long time and still be shitty.

I think it's prevailed for all of human history in this case, so while shitty it appears to be a feature

Re: Training our own AI models

#154

Post by skrebbel »

deflator wrote: post #69
skrebbel wrote: post #53

Earlier quoted context omitted.

They could’ve done a lot worse, and most companies would’ve.

There's a helpful response. It could always be worse!

Okok what I meant was, too many companies would've hid this in a ToS update, and these guys actively market the change, and the worst crime we can accuse them of is using the wrong word for "enable". I agree that's not perfect, but I can totally imagine that if you use "opt-in" and "opt-out" a bit too much at work, esp as an analytics company deep in cookie law land, that "opted-in be default" sounds like a perfectly normal thing to say, ie I bet this is all not malice, just mild incompetence.

Re: Training our own AI models

#155

Post by abustamam »

micromacrofoot wrote: post #153
abustamam wrote: post #136

Earlier quoted context omitted.

Two things can be true. A shitty thing can prevail for a long time and still be shitty.

I think it's prevailed for all of human history in this case, so while shitty it appears to be a feature

Agreed but "it's always been shitty" isn't a reason to do nothing about it.

For much of human history slavery was fine too, but thankfully that's been dealt with in many nations. I know some folks argue that capitalism is just the next iteration of slavery but that's different.

Re: Training our own AI models

#156

Post by micromacrofoot »

abustamam wrote: post #155
micromacrofoot wrote: post #153

Earlier quoted context omitted.

I think it's prevailed for all of human history in this case, so while shitty it appears to be a feature

Agreed but "it's always been shitty" isn't a reason to do nothing about it. For much of human history slavery was fine too, but thankfully that's been dealt with in many nations. I know some folks argue that capitalism is just the next iteration of slavery but that's different.

not sure why you think stating facts implies nothing should be done about it

slavery is still fine in the US as long as you're incarcerated, and the US still does business with countries that regularly use the old fashioned kind — hell, some people still want to bring chattel slavery back here and still proudly display the confederate flag

these are just very slow things to change... in the timespan of centuries if you're lucky, but in the case of wealth bringing privilege we haven't cracked the issue for thousands of years

Post reply on HN