discountopinion

Yeah agreed and this is the clincher.. providing above as-is experience will cost a tonne in inference and thus Amazon may bow out as cannot see the business case. Apple on the other hand as a you and I mentioned can do PCC based queries for old HomePods and either go augmented local SoC for better experience or remain full cloud. The trend is that Apple will favour on device LLMs as CapEx is borne by the customer and given the likely SoC evolution of Apple Watch then future HomePods may have a strong local slim LLM processing capability in addition to PCC offloading.

It does not seem to be too expensive to add the beefed up NPU to the next SoC for AW and HomePods to deliver 38+ TOPS locally. HomePod AI edition for the win.

In summary Apple's AI roll out is super smart, business aligned and CapEx light. I am a huge fan of the HomePod format of compute and I really home Apple doubles down on its value prop.

roundaboutnow said:

discountopinion said:

Hi Dewme,

I see your points and my points were not articulated well.

Amazon seems to be abandoning Alexa and giving up on the ecosystem since it was an initial loss leader for a way to shop with Amazon that never appeared. Amazon will not continue to subsidize the Alexa ecosystem forever.

the current use cases for both Siri and Alexa merely scratches the surface of what voice based computing can offer. Children in particular love having conversations and learning. People with vision impairment benefit from HomePods a lot.

i didn’t mean for apple to aim for the price point of the cheap echo dots, I meant for apple to lean in on device local ai above and beyond what Alexa and current Siri are capable of.

As home automation becomes ever more intrusive in our lives then apples privacy ethos is a great selling point. Amazon is terrible.

the cost and price point will be determined by the capabilities of the next Apple Watch Ultra soc.

dewme said:

discountopinion said:

Ideally Apple will want to do a platform play and invest to also kick Alexa in the teeth. The Echo Dots of this world would then be painfully obsolete. If Apple can do a software only upgrade of AI capability through some magic then they will win the internet. There is a slight chance they can make it happen as current Siri models have a larger proportion of on device processing.

The Echo Dots are sub $50 USD devices that get heavily discounted during Amazon’s sales events. I don’t see Apple targeting anything in that price range and sales volume. They are ubiquitous, cheap, work well, have voice recognition that puts Siri to shame, and could probably be sold in the checkout line in grocery stores. I use mine for turning on/off smart lights and smart plugs, in-house intercom, getting weather updates, setting alarms and timers, monitoring interior temperature, and automation. None of these things require more horsepower than what’s already in them. Some of the other more advanced capabilities of the Echo Dot, including integration with alarm systems using sound recognition, serving as mesh points on certain brand WiFi systems, and in the larger models, halfway decent audio playback.

Even the cheapest and oldest Dots are more than enough for listening to radio quality music, sports broadcasts, talk shows, podcasts, etc., especially while working in the garage or shed. I wouldn’t dare put an HomePod in my garage due to living in a fairly challenging 4-season climate with snow, rain, cold, high humidity, and dust always present, regardless of the much higher price and lower functionality of Apple’s smart speakers.  Even older Echo Dot models have been updated to support Thread and Matter. How many older Apple devices are being backfitted with Thread and Matter support? 

There’s no doubt that a 20-ton excavator would kick in the teeth of a hand shovel. But what’s the point? The Echo Dot and pretty much every current Echo device is way too low of a bar for Apple to aim for. Apple also needs to figure out how to monetize AI on some of these devices. Unlike Amazon, Apple has far more options available to take advantage of, like its customer centric ecosystem. Apple has clearly set its targets appropriately and is executing on a plan that is at least a decade in process. 

Obviously, the Echo Dots rely on server based processing to provide what voice capabilities they have. One possible solution for the HomePods and Apple TVs would be to give the customer the ability to opt-in to Apple's Private Cloud in a manner similar to what is being proposed for access to ChatGPT when a request exceeds Apple Intelligence on-device capability.

Hi Dewme,

I see your points and my points were not articulated well.

Amazon seems to be abandoning Alexa and giving up on the ecosystem since it was an initial loss leader for a way to shop with Amazon that never appeared. Amazon will not continue to subsidize the Alexa ecosystem forever.

the current use cases for both Siri and Alexa merely scratches the surface of what voice based computing can offer. Children in particular love having conversations and learning. People with vision impairment benefit from HomePods a lot.

i didn’t mean for apple to aim for the price point of the cheap echo dots, I meant for apple to lean in on device local ai above and beyond what Alexa and current Siri are capable of.

As home automation becomes ever more intrusive in our lives then apples privacy ethos is a great selling point. Amazon is terrible.

the cost and price point will be determined by the capabilities of the next Apple Watch Ultra soc.

dewme said:

discountopinion said:

Ideally Apple will want to do a platform play and invest to also kick Alexa in the teeth. The Echo Dots of this world would then be painfully obsolete. If Apple can do a software only upgrade of AI capability through some magic then they will win the internet. There is a slight chance they can make it happen as current Siri models have a larger proportion of on device processing.

The Echo Dots are sub $50 USD devices that get heavily discounted during Amazon’s sales events. I don’t see Apple targeting anything in that price range and sales volume. They are ubiquitous, cheap, work well, have voice recognition that puts Siri to shame, and could probably be sold in the checkout line in grocery stores. I use mine for turning on/off smart lights and smart plugs, in-house intercom, getting weather updates, setting alarms and timers, monitoring interior temperature, and automation. None of these things require more horsepower than what’s already in them. Some of the other more advanced capabilities of the Echo Dot, including integration with alarm systems using sound recognition, serving as mesh points on certain brand WiFi systems, and in the larger models, halfway decent audio playback.

Even the cheapest and oldest Dots are more than enough for listening to radio quality music, sports broadcasts, talk shows, podcasts, etc., especially while working in the garage or shed. I wouldn’t dare put an HomePod in my garage due to living in a fairly challenging 4-season climate with snow, rain, cold, high humidity, and dust always present, regardless of the much higher price and lower functionality of Apple’s smart speakers.  Even older Echo Dot models have been updated to support Thread and Matter. How many older Apple devices are being backfitted with Thread and Matter support? 

There’s no doubt that a 20-ton excavator would kick in the teeth of a hand shovel. But what’s the point? The Echo Dot and pretty much every current Echo device is way too low of a bar for Apple to aim for. Apple also needs to figure out how to monetize AI on some of these devices. Unlike Amazon, Apple has far more options available to take advantage of, like its customer centric ecosystem. Apple has clearly set its targets appropriately and is executing on a plan that is at least a decade in process.