"The WeChat team went ahead and plugged in an external open-weight rival, DeepSeek R1, because it was better than anything Yuanbao had made."
It's hard to understand why Apple isn't willing to do the same by taking an open-source LLM, modifying and customizing it to suit their needs, and plugging it into Siri to instantly make it far better. Using a customized open model on Apple's own servers would allow it to keep its core commitment of privacy while avoiding the need to rely heavily on OpenAI or Anthropic or to reinvent the wheel by building its own models.
That's definitely a really interesting take. If they can integrate a strong open-source LLM into their products, the stickiness will definitely make it hard for people to want to switch. I guess it's a matter of negotiations for how the profit split looks then? Maybe they're in talks with OpenAI (seems like they're embracing more bigtech partnerships now) in the background already!
I wonder if Apple has the right talent to productize the model. Hard to imagine they don’t, but all evidence suggests not.
The only other thing I can imagine is that Apple has become too siloed because of their legendary internal secrecy. They may have locked down interdepartmental info so hard that they’ve lost their systems thinking mojo.
I do wonder though: Is OpenAI’s recent turn toward agentic offering an acknowledgment that China’s app-centered strategy is ultimately the better one across all markets?
It’s almost like OpenAI is retrofitting itself to make up for billions wasted on a science experiment with no end.
Put another way: When ever was “destiny” a real strategy. (Never)
Perhaps, looking two steps ahead, we’re witnessing convergent evolution.
And if so, maybe China really is ahead once the real international AI race does begin.
The sputnik has been launched, but the astronaut thinks she’s just driving an SUV. ;)
I think we're also just still in such an early stage of AI development that it's tough to say which company or country has "won".. it will be interesting to see how this industry plays out in the long run. I mean, think about the OG internet players like Yahoo or eBay ... their relevance deteriorated over the years...
"The WeChat team went ahead and plugged in an external open-weight rival, DeepSeek R1, because it was better than anything Yuanbao had made."
It's hard to understand why Apple isn't willing to do the same by taking an open-source LLM, modifying and customizing it to suit their needs, and plugging it into Siri to instantly make it far better. Using a customized open model on Apple's own servers would allow it to keep its core commitment of privacy while avoiding the need to rely heavily on OpenAI or Anthropic or to reinvent the wheel by building its own models.
That's definitely a really interesting take. If they can integrate a strong open-source LLM into their products, the stickiness will definitely make it hard for people to want to switch. I guess it's a matter of negotiations for how the profit split looks then? Maybe they're in talks with OpenAI (seems like they're embracing more bigtech partnerships now) in the background already!
I wonder if Apple has the right talent to productize the model. Hard to imagine they don’t, but all evidence suggests not.
The only other thing I can imagine is that Apple has become too siloed because of their legendary internal secrecy. They may have locked down interdepartmental info so hard that they’ve lost their systems thinking mojo.
It’s a great question though
i think for Llama, this is not truly free for a company the size of Apple.
Llama restricts use by large scale corporations, but DeepSeek R1 and Kimi both use highly permissive licenses that allow commercial use.
I love this insight.
I do wonder though: Is OpenAI’s recent turn toward agentic offering an acknowledgment that China’s app-centered strategy is ultimately the better one across all markets?
It’s almost like OpenAI is retrofitting itself to make up for billions wasted on a science experiment with no end.
Put another way: When ever was “destiny” a real strategy. (Never)
Perhaps, looking two steps ahead, we’re witnessing convergent evolution.
And if so, maybe China really is ahead once the real international AI race does begin.
The sputnik has been launched, but the astronaut thinks she’s just driving an SUV. ;)
I think we're also just still in such an early stage of AI development that it's tough to say which company or country has "won".. it will be interesting to see how this industry plays out in the long run. I mean, think about the OG internet players like Yahoo or eBay ... their relevance deteriorated over the years...
we can steal from them-turnabout is fair play
haha, if it's all open- how's that stealing?
good point,grace-let's call it 'borrowing' or 'using'