The only way you could catch up faster is by having orders of magnitude more hardware. But due the global shortage on hardware right now, nobody has that, not even Google.
The only way you could catch up faster is by having orders of magnitude more hardware. But due the global shortage on hardware right now, nobody has that, not even Google.
This reads like wishful thinking, in the absence of a real moat. Do you think the hyperscalers who are buiding data centers as quickly as they can get permits do not have the infra scale to match OpenAI? Who do you suppose has been providing ever-increasing revenue to Nvidia these past few years? They didn't start doing ML experimentation when ChatGPT was announced.
I agree that they aren't specifically giving LLMs all of their attention - their infrastructure still has to support existing, profitable products that bring in real revenue. I disagree that they can't experiment to the same scale.
Google has nearly unlimited resources to pursue this shit.