pineapple_pizza@lemmy.dexlit.xyz
on 26 Oct 2023 05:03
nextcollapse
We’re in the sliceline era of generative ai, enjoy it before prices get hiked
nicetriangle@kbin.social
on 26 Oct 2023 09:20
collapse
Yep everyone's trying to capture market share and stamp out any competitors with shorter funding runways until they achieve some amount of monopolization over the customer base. Then comes the price hikes and other anti consumer bullshit.
That’s not such a big deal. Their objective is to get people hooked on the system. After that, they’ll jack up the price. Microsoft can easily afford to lose money for several years in pursuit of that target.
(One way this plan could fall through is if LLM tech progresses to the extent that free and open source copilots, run locally, can give result that are just as good.)
theterrasque@infosec.pub
on 26 Oct 2023 06:01
nextcollapse
There are already very impressive local models for coding. Some have come out favourably to copilot in tests iirc
Pechente@feddit.de
on 26 Oct 2023 06:03
nextcollapse
One way this plan could fall through is if LLM tech progresses to the extent that free and open source copilots, run locally, can give result that are just as good.
Performance is not great but apparently it’s not optimized at all as of right now.
HidingCat@kbin.social
on 26 Oct 2023 06:51
collapse
Not familiar with the tech, but wouldn't server-side LLMs still have an advantage regardless because of the greater power available on tap? Anything that improves local LLM will also benefit server-side LLMs, wouldn't it?
my_hat_stinks@programming.dev
on 26 Oct 2023 07:44
nextcollapse
Not necessarily, as it gets faster the latency between your local and remote machines becomes a bigger fraction of the time taken to process anything. If your local machine processes in 50ms and the remote machine in 5s, a latency of just 45ms would make your machine faster.
Running locally also cuts out a lot of potential security issues inherent to sending data over a network, and not sending your data to a third party is a bonus too.
Possibly, but given the choice between paying $20/m for a marginally better version of something that’s free and probably built in to your editor at that point, most people would probably take the free thing. At that point paid llms will need to find new niches beyond simply existing.
worldsayshi@lemmy.world
on 26 Oct 2023 07:46
collapse
That seems so weird when you think about the pricing for openai API. It feels at least an order of magnitude cheaper than using chatgpt plus subscription, which in turn is $20/month. If Copilot is losing money, openai must be burning money by truckloads.
threaded - newest
…and might be losing money from each of those: neowin.net/…/microsoft-reportedly-is-losing-lots-…
We’re in the sliceline era of generative ai, enjoy it before prices get hiked
Yep everyone's trying to capture market share and stamp out any competitors with shorter funding runways until they achieve some amount of monopolization over the customer base. Then comes the price hikes and other anti consumer bullshit.
That’s not such a big deal. Their objective is to get people hooked on the system. After that, they’ll jack up the price. Microsoft can easily afford to lose money for several years in pursuit of that target.
(One way this plan could fall through is if LLM tech progresses to the extent that free and open source copilots, run locally, can give result that are just as good.)
There are already very impressive local models for coding. Some have come out favourably to copilot in tests iirc
Edit: evalplus.github.io/leaderboard.html
MS might be in trouble then.
Performance is not great but apparently it’s not optimized at all as of right now.
Not familiar with the tech, but wouldn't server-side LLMs still have an advantage regardless because of the greater power available on tap? Anything that improves local LLM will also benefit server-side LLMs, wouldn't it?
Not necessarily, as it gets faster the latency between your local and remote machines becomes a bigger fraction of the time taken to process anything. If your local machine processes in 50ms and the remote machine in 5s, a latency of just 45ms would make your machine faster.
Running locally also cuts out a lot of potential security issues inherent to sending data over a network, and not sending your data to a third party is a bonus too.
Possibly, but given the choice between paying $20/m for a marginally better version of something that’s free and probably built in to your editor at that point, most people would probably take the free thing. At that point paid llms will need to find new niches beyond simply existing.
That seems so weird when you think about the pricing for openai API. It feels at least an order of magnitude cheaper than using chatgpt plus subscription, which in turn is $20/month. If Copilot is losing money, openai must be burning money by truckloads.
It sure is, what would make you think otherwise? It has enough VC money to burn
Openai is losing money. I’m too lazy to find the article, but it is mindblowing how much they’re losing
Edit: nvm, I found it medium.com/…/openai-lost-540m-in-2022-needs-100b-…
Those cost seems to be developmental and not operational, after they reach AGI I assume it will go down
After they reach AGI they will own the entire world
Yeah, I'm sure Microsoft is happy with the theft of copyrighted works and people's personal information.