Live data from Hacker News

Training our own AI models

posthog.com

91–100 of 156 posts

Re: Training our own AI models

#91

Earlier quoted context omitted.

this is the business model of all companies training AI, if they had to get permission we wouldn't have frontier LLMs at all

So it's OK to do stuff without permission as long as we get something that makes a lot of people a lot of money?

*"that makes a select few rich people a lot more money?"

Re: Training our own AI models

#92
post #4

Most companies would bury this change in a deceptively boring T&Cs update, but we value transparency, so here's what you need to know in an internet-friendly numbered list: Users on our EU cloud instance are opted out by default So too users with agreements that prevent training (e.g. BAA, MSA, or similar) All other users on our US cloud instance are opted in by default We will anonymize all data before it's used for…

> We will do all the model training ourselves

That's actually an interesting note. So you all will be managing the training runs on hardware you own or rent and manage?

Re: Training our own AI models

#93
post #27

Earlier quoted context omitted.

Which we probably need to consider changing now that some truly bizarre and evil shit is being done on donor organs: https://news.ycombinator.com/item?id=48212992

Off topic but I'm curious what's evil about that article. Bizarre, sure, but evil? They took a brain (an organ) from a dead person who had previously given consent to use their organs for that purpose.

I think most people, and definitely myself, when they sign up to be an organ donor, they think of the liver, kidney, and heart. Things that will help other people.

Not the brain, the center of conscience, being kept in some sort of horror-movie half-alive state. I do not think we understand consciousness enough to rule out what those brains are experiencing.

Re: Training our own AI models

#94
They say, "our goal here is to improve PostHog as a product for our customers, not to expose or sell models trained on your data" but then don't actually list that as a limitation in the bulleted points.

AFAICT this now gives them default permission to train an LLM on your code (as Posthog telemetry data is inextricably tied to your code) use it, and even sell it if they wanted to (as it's not your data anymore, it's their model). Yikes.

Re: Training our own AI models

#95

Earlier quoted context omitted.

this is the business model of all companies training AI, if they had to get permission we wouldn't have frontier LLMs at all

So it's OK to do stuff without permission as long as we get something that makes a lot of people a lot of money?

It's not OK, but it's a fact.

Re: Training our own AI models

#96
post #57

Gross. They’ll use your product and your data to later sell a product back to you.

Even if there were no AI, that's not any different than any SaaS where your data gets stored. Picking at random, Optimizely certainly has a ton of interaction data available and they build new features and products that leverage your data (without which the features would be impossible). Could be reporting tools, funnel analysis, etc.

Re: Training our own AI models

#98
I’ve been evaluating PostHog for our company.

I’ve now made our decision. We won’t be using them.

If they are going to position yourself as the non-slimy no-BS guys, they can’t pull this nonsense.

Re: Training our own AI models

#99
post #4

Most companies would bury this change in a deceptively boring T&Cs update, but we value transparency, so here's what you need to know in an internet-friendly numbered list: Users on our EU cloud instance are opted out by default So too users with agreements that prevent training (e.g. BAA, MSA, or similar) All other users on our US cloud instance are opted in by default We will anonymize all data before it's used for…

There is definitely some confusion on the EU part. I am a European citizen, but some of my activity data on some of the sites I host is logged in US Posthog, which means Posthog is subject to the GDPR, even if the data is US hosted!
Post reply on HN