18
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 13 Oct 2024
18 points (100.0% liked)
TechTakes
1489 readers
30 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
NYT: "Microsoft and OpenAI’s Close Partnership Shows Signs of Fraying" (https://www.nytimes.com/2024/10/17/technology/microsoft-openai-partnership-deal.html)
Archive: https://archive.ph/Bas23
HN link from whence I got this, have not mined this for sneers: https://news.ycombinator.com/item?id=41878281
edit oh wait found one
Can't enshittify that which is already shit
Just a normal hackernews, testing if the models they use are racist
The unspecificity is damning. "Facts about their membership" might range from "what racial separatist group is Skum Shitt (R, NC) a former member of" to "am I eligible to join The Brotherhood of Untarnished Ejaculate".
That's an interesting example to pair up with the one about racist hate groups. Unconventional in what way, motherfucker?
tl;dr of the article: ever since the ousting of altman, microsoft, which virtually owns openai, has been suspicious of openai's actual worth. therefore MS has cut down on the infinite resource flow. openai employees are whining about this.
there is one additional point in three of the near final paragraphs, which I'll quote in full because they are so amusing to me
The first paragraph surprised me. I didn't know there were still some true believers left.
I think I see a possible future here. Just as the promptfondlers are now trying to talk down human accomplishments to make the LLMs sound more impressive ('it learns just like a child!' (no, it doesn't)). As when you are trying to reach a deal on a car, you either need the buyer to raise the price, or the seller to lower theirs. This will lead to a lawsuit where they are going to drop down the theoretical capabilities of an AGI just to trigger this clause.
And as the judge thinks that emoji's are a form of novelty pasta, any potential jury can't spell stattistical, there is a 50% chance that they will be convinced it is AGI because humans like the AGI also make mistakes.
Still unclear how OpenAI declaring they've created an AGI (and getting this adjudicated as "true") and thus being released from their MSFT partnership helps OpenAI. They're just lost all their compute, and will have a worse negotiating position from which to get more compute. What's the upside?
Yeah fair
They can sell out to someone else instead? Amazon seems to want to be an AI company, for example, and their current offering isn’t great even by the relaxed standards of LLMs.