9 ms·
This is pretty much to go head-to-head with Apple (and Samsung). Both are making leaps and strides using “neural coprocessors” and the like for running models o
by synaesthesisx 3y ago
This is pretty much to go head-to-head with Apple (and Samsung). Both are making leaps and strides using “neural coprocessors” and the like for running models on mobile hardware. Mobile edge computing is where we’re going to see a lot of use cases that enable functionality while maintaining data privacy & performance.
Keep in mind “mobile devices” extends past just smartphones, onto wearables/headsets as well.
- vmurthy 3y ago> enable functionality while maintaining data privacy Ah but is it in (insert any ad business big-tech company)'s best interests to do this? Do they already have so much data that these user <> LLM interactions are only marginal that they don't care about harvesting the data ?
- dan-robertson 3y agoFundamentally, if they think their users want this enough (for it to be profitable), they should be willing to give their customers what they want. Seems reasonably compatible with Apple’s business, for example.