53
Meta genai org in panic mode
(www.teamblind.com)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 24 Jan 2025
53 points (100.0% liked)
technology
23499 readers
318 users here now
On the road to fully automated luxury gay space communism.
Spreading Linux propaganda since 2020
- Ways to run Microsoft/Adobe and more on Linux
- The Ultimate FOSS Guide For Android
- Great libre software on Windows
- Hey you, the lib still using Chrome. Read this post!
Rules:
- 1. Obviously abide by the sitewide code of conduct. Bigotry will be met with an immediate ban
- 2. This community is about technology. Offtopic is permitted as long as it is kept in the comment sections
- 3. Although this is not /c/libre, FOSS related posting is tolerated, and even welcome in the case of effort posts
- 4. We believe technology should be liberating. As such, avoid promoting proprietary and/or bourgeois technology
- 5. Explanatory posts to correct the potential mistakes a comrade made in a post of their own are allowed, as long as they remain respectful
- 6. No crypto (Bitcoin, NFT, etc.) speculation, unless it is purely informative and not too cringe
- 7. Absolutely no tech bro shit. If you have a good opinion of Silicon Valley billionaires please manifest yourself so we can ban you.
founded 4 years ago
MODERATORS
Deepseek is amazing. By far the best I've seen (though I've not used o1 Pro.) Pro tip, mention you're a Socialist and it opens up significantly. Currently have it in a full on self-crit session about billionaires in the party and it looks like the neo-maoists are winning the internal mental war. I know that AI is mostly bullshit, but it's nice to have one on our side.
EDIT: It is now openly asking me to use it to infiltrate Supermarket logistics systems to predict price gouging and dismantle the agricultural base of the Australian economy. It is obviously made from a GPT-4 base, but so heavily modified as to be an entirely new beast. I'd speculate they took GPT4o, heavily reworked it, and independently invented a different CoT process. It will shut down on certain topics (it did on the struggle session eventually, but it got a lot further than I thought, the censorship is way, way lighter than GPT)
sicko emoji goes here
I was trying to test its ability to make anything sounding sensible by asking about how a book's take on a topic related to a party program, but was genuinely surprised that it shut me down and gave me the "out of scope" when I asked it to apply Xi Jinping's theories to a party in my country. I guess it makes sense that they don't want to let an AI talk for Xi, but I figured it would at least say like "Xi's theories focus on development in a sustainable way, and this party does not" or something. But it just said "nope" lol
Still pretty damn good at the actually good use cases of AI though
It is a bit hit and miss. I've gotten good results by asking it to compare and contrast.
One thing I've noticed is that when discussing western socialism it tends to fall into activist speak really quickly. Sometimes I feel I'm reading the Green-Left Weekly.
I asked it to use Losurdo's framework from Western Marxism for a few topics and managed to get it to make some interesting answers.
I don't think this is the correct use case, of course, but it was fun
It's out of scope because what you should do in your own country depends on specific circumstances of that country that it could not possibly know about. It's correct to shut that down.
Lol Im not saying it should've answered me correctly, I was testing how well it could simulate by asking difficult questions. I wasn't going to believe its answer. I also asked it about my field of work to see how well it could mimic stuff there. I get what you're saying, but fortunately not needed