I think it's about convention and adoption - "first taste is free" to get as many ppl as possible accustomed (read: addicted) to using GenAI. Plus most majors are currently releasing open weight models.
...but i suspect few to none will continue releasing free and open models when they start cutting too deeply into token profits.
We'll likely see smaller open models, but few who have the resources to train large, frontier sized models will want to do so for free... maybe China, but they're also a dubious choice for many.
It's a much cheaper way to run a freemium pricing plan than to give out actual time on a real GPU.
The free models aren't as advanced as the commercial ones, so you still need to pay to get better service. Many companies do this by hosting the smaller models for free, and charging for access to the bigger ones, but the costs of running even the small models adds up, so if they can get you to run it on your own hardware, they save a lot on GPU time, but still get the advertising exposure of a freemium pricing plan.
Giving away or offering lower cost AI models help keep users in the Google ecosystem, which is very financially lucrative for Google with its ad sales related to its search engine.
It keeps "branding" intact. This is similar to red-bull spending $3 BILLION annually on marketing... it's to keep the brand alive in the minds of folks and "associate" them as kinda the de-facto in given industry.
All it's done is make me scratch my head and wonder how many millions of dollars was spent to develop something and then just give it away with no strings.
I'm sure there are strings somewhere but I'm not sure what they lead to.
I haven't felt any inclination to utilize their cloud models because of it.
I think it's about convention and adoption - "first taste is free" to get as many ppl as possible accustomed (read: addicted) to using GenAI. Plus most majors are currently releasing open weight models.
...but i suspect few to none will continue releasing free and open models when they start cutting too deeply into token profits.
We'll likely see smaller open models, but few who have the resources to train large, frontier sized models will want to do so for free... maybe China, but they're also a dubious choice for many.
It's a much cheaper way to run a freemium pricing plan than to give out actual time on a real GPU.
The free models aren't as advanced as the commercial ones, so you still need to pay to get better service. Many companies do this by hosting the smaller models for free, and charging for access to the bigger ones, but the costs of running even the small models adds up, so if they can get you to run it on your own hardware, they save a lot on GPU time, but still get the advertising exposure of a freemium pricing plan.
Giving away or offering lower cost AI models help keep users in the Google ecosystem, which is very financially lucrative for Google with its ad sales related to its search engine.
You don't need the google ecosystem at all to utilize them though how does this keep me there whatsoever?
It keeps "branding" intact. This is similar to red-bull spending $3 BILLION annually on marketing... it's to keep the brand alive in the minds of folks and "associate" them as kinda the de-facto in given industry.
I'm assuming the bet is that giving away the local models encourages developers to use their cloud ones for more complex tasks
All it's done is make me scratch my head and wonder how many millions of dollars was spent to develop something and then just give it away with no strings.
I'm sure there are strings somewhere but I'm not sure what they lead to.
I haven't felt any inclination to utilize their cloud models because of it.
Speculation: It helps engineer consumer agency.