this post was submitted on 22 Jun 2023
130 points (100.0% liked)
Technology
37747 readers
311 users here now
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
Subcommunities on Beehaw:
This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I suspect they have signed an exclusivity deal with some kind of third party to use the API. It could be for "AI" or it could be for more nefarious purposes.
Spez knows he can create 'traffic' of user comments and answers with AI. He also knows he can use AI to moderate subreddits. He doesn't care about the quality of the site, just the numbers that get him his payday. He'll burn it to the ground and cash-out, leaving a mess in his wake.
OpenAI CEO Sam Altman sat on the reddit board for years and was briefly CEO for 8 days.
OpenAI is another dumpster fire waiting to happen, then.
Why else would they make access to OpenAI/ChatGPT/etc so cheap? So others can build businesses on the tech that get locked in before they jack up the price.
We've seen this rodeo plenty of times now.
I don't disagree with your saying that is why they're doing it - but I wonder how well it will actually work out for them. Natural Language Prompting is hard to "lock in" someone on. Sure, the complex jobs with custom trained models are going to get locked in for sure - but the companies that are just adding "chat bots" to their apps? I don't see the difficulty in migration.
I use OpenAI for one of my projects, and frankly there's little that would keep me from being able to migrate to another service if one came along that gave a better value. An AI platform isn't like an IaaS platform, there really isn't a lot of platform-specific workflows involved, and prompts that work on one LLM should work just about as well on another.
Even for custom trained models, most training data is stored in json, and should be easy to feed into another LLM, though of course tweaking will be required
That's why it's important to go back thru our comment history and replace them with linguistic garbage. To ensure Reddit can't profit off our donations. I'm not in the business of subsidizing Reddit, after all.
"Plonked up behind the radio them ready the plastic manuscript who observe Jerry's can." Or whatever.
If I were implementing this nefarious Reddit I probably wouldn't have edits wipe out the original Tatars data. It's certainly not necessary to implement edits that way.
In fact, the editing log itself can be used as more data.
We actually know for a fact they don't do it that way, since Reddit has already been caught undoing peoples "delete" edits after they've gone