3 ms·

I never said anything about openai not being relevant. the point is scale and good enough. gpt-4 is already good enough for most reasonable use cases, notably s

I never said anything about openai not being relevant. the point is scale and good enough. gpt-4 is already good enough for most reasonable use cases, notably siri optimization. 5 years is a long time. 5 years ago llms didn't even exist.

shmoogy 4y ago
Gpt3.5-turbo runs laps around siri and is cheap, surely Apple will acquire talent to build something that does remote queries and falls back to local for the next 1-2 years while they figure out on device accelerated LLMs
- scarface74 4y ago
  People keep saying this. But it’s only half the problem to process natural language in text form. How well would ChatGPT do at understanding Spanish from a non native speaker with a southern US accent? (Raises hand). I’ve yet to see a speech to text system that does well at understanding speech by non native speakers or even native English speakers with a heavy regional accent.
  - yazaddaruvala 4y ago
    We have yet to see any Large Speech Models or Large Multimodal Models that use speech and text as inputs. It’s only the being but the general process to build this stuff now exists and the value proposition is crystal clear.
- crooked-v 4y ago
  I don't think they'd even need to go that far to get started. Alpaca-LoRA is better than Siri when it comes to understanding queries and can already run on consumer GPUs. I would be really surprised if they couldn't get something equivalent running on the top-series A-chip devices in relatively short order.
smoldesu 4y ago
> 5 years ago llms didn't even exist Almost 5 years ago TalkToTransformer did 80% of ChatGPT's job with 0% of the hype. Once people realize just how glacial all this stuff moves, I think the honeymoon phase will be over.
- endisneigh 4y ago
  right, that's why I mentioned performance at the current level at far less compute, and not necessarily better performance.

shmoogy 4y ago

Gpt3.5-turbo runs laps around siri and is cheap, surely Apple will acquire talent to build something that does remote queries and falls back to local for the next 1-2 years while they figure out on device accelerated LLMs

scarface74 4y ago
People keep saying this. But it’s only half the problem to process natural language in text form. How well would ChatGPT do at understanding Spanish from a non native speaker with a southern US accent? (Raises hand). I’ve yet to see a speech to text system that does well at understanding speech by non native speakers or even native English speakers with a heavy regional accent.
- yazaddaruvala 4y ago
  We have yet to see any Large Speech Models or Large Multimodal Models that use speech and text as inputs. It’s only the being but the general process to build this stuff now exists and the value proposition is crystal clear.
crooked-v 4y ago
I don't think they'd even need to go that far to get started. Alpaca-LoRA is better than Siri when it comes to understanding queries and can already run on consumer GPUs. I would be really surprised if they couldn't get something equivalent running on the top-series A-chip devices in relatively short order.

smoldesu 4y ago

> 5 years ago llms didn't even exist Almost 5 years ago TalkToTransformer did 80% of ChatGPT's job with 0% of the hype. Once people realize just how glacial all this stuff moves, I think the honeymoon phase will be over.

endisneigh 4y ago
right, that's why I mentioned performance at the current level at far less compute, and not necessarily better performance.