The way, financially, for a company to win the ChatBot game in a profitability sense is—probably—not to play. Cupertino’s AI contrarianism of betting on the device, not the cloud; and on the...
I wonder at Microsoft’s approach. It seems better for Microsoft to concentrate on business apps and avoid consumer chat bots. They don’t do well in the consumer market.
Cloud service is nice to have. It makes the filing system more like PC’s. It is possible to duplicate files on Icloud. I don’t use upload or download anymore, just copy and paste from IPad to Icloud.
The default browser for this IPad 10 is Google, and Google Ai creates the Ai information. I can’t see why Apple would have to have their own, on top of what Google can find.
I think Apple is trying to price some of their hardware at lower levels. There is an Iphone at Walmart. This IPad was sold at Costco during Christmas season, mostly for gifts, to push people who don’t have experience with Apple into using Apple.
Apple is hoping people will buy services. I did get a cloud subscription shortly after I received this gift.
Apple spreadsheet Numbers works well, but it is a struggle sometimes to find out how to do things. It is kind of hard to work with the 5 column limit, but I got used to it. I used the drawing tool on some spreadsheets, which is a very nice feature. Not sure if the drawings will transfer to PC’s.
What five column limit? Numbers spreadsheets have any number of columns. Just drag the sheet horizontally. The interfaces is a bit clunky compared to VisiCalc on an Apple II but good for the new century.
That sounds like a five column limit for the column outline structure. Do they also have a five row limit? They probably consider five levels deep enough for an outline as each added column multiplies the columns in the displayed table.
Then again, I’ve never used pivot tables, but I did translate Apple II Visicalc to run on a Z80 Trash 80.
I saw that 5 column limit when reading about pivot tables. I wouldn’t try 1000 columns and 1 million rows without other information at about what file size limit I should use with this small IPad.
Five columns is about all I can see when there are larger column widths. If I wanted to use the data for a pivot table, it would not be necessary to re-do the table.
Where I saw this 5 column limit was in reading about pivot tables.
Here is a copy of an Ai message. My search bar: “is there a 5 column limit for a pivot table in Numbers IPad 10.
Response:
“According to users in the Apple Support Communities, Numbers on the iPad has a limit of 5 fields in the "Columns" and "Rows" boxes in the Pivot Options for Pivot Tables.”
I started using small computers when they first came out, early 1980’s. Visicalc was the first spreadsheet program, for the first IBM machine, predating Lotus 123 and Excel. I wonder if they just decided to use the same name, what a surprise.
Earnings look great for the Magnificent 6 (Tesla is a train wreck in an earthquake). But depreciation spreads out cap-ex for years to come; it will be a weight on future earnings if future AI revenue doesn't come. Also, the OBB Bill accelerates cap-ex tax expenses, so lower taxes boosts current earnings too. Everyone else's depreciation lagged expenses shows up as today's revenue for NVIDIA.
We can see it in the yawning gap between some earnings and free cash flows. AAPL and MSFT are exceptions in the most recent quarter ($ billions). Why shouldn't Apple rent someone else's AI, or let the user choose similar to search engines? They book the cash, others pay the expenses.
AAPL Net Income: $23B; Free Cash Flow: $24B
MSFT NI: 27; FCF: 26
NVDIA NI: 19; FCF: 26 (prior quarter balances this out)
A key point is that Apple is focusing on narrow AI for particular tasks. This echoes the GOFAI era of Roger Shanks' "scripts". It is far easier to reduce ambiguity when you know how to act and respond. Porting this approach to a private edge device makes a lot of sense, and is akin to my earlier comments that AI needs to be more niche and less general to be profitable.
Secondly, should there be an AI bust, those NVIDIA chips will sell at auction for cents on teh dollar. Apple can then snap up all it wants for its more generalized cloud AI.
But the critique of not delivering useful AI stands. Apple has failed to deliver production versions of its AI on devices to reliably achieve those narrow tasks. That is a major failure that needs to be fixed.
A friendly comment. I read you EFD and admire your scope and clarity. This posting although thoughtful, cogent, and as far as I can see (as a non apple person) correct, is (with all due respect) repetitive. Especially in your description of the cost of the arms race, and the Nvidia rent. I suspect you might be whelmed by the conflagration of Benjamins.
Thank you❣️I am starting to look at a new computer, I’ve always had apple products. Due to the current yet ridiculous word police having privacy is very important as well. I encourage people to be regular listeners to your site or better yet support it! as I do. thanks again.
Very well explained and appreciated. I definitely agree that people will be focused on their phones for most of what they will use AI for in terms of non-business-related activities and likely to use their PCs or Macs for business planning, etc.
Apple started offering voicemail transcripts two or three years ago. Since I can’t usually understand recorded speech, this was immensely useful. It’s quicker, too. Apple’s Magnifier app has a mode to speak or display what the phone camera sees. They advertised it as an aid for the visually impaired, but it is hard to miss an entire business area. In actually providing useful AI, Apple does fine. Anyone who relies on a big LLM to summarize things is setting for a fall.
The AI push is essentially about undoing the PC rebellion and moving back to the 20th century model of centralized control. As the Chinese and Apple have shown, it is possible to provide useful AI at the periphery. The Chinese LLM performed AI at a moderately high level but with extremely low power consumption. Think VW Beetle in 1963. Apple is performing AI on wrist watches. Meanwhile Detroit is offering bigger, more powerful, unreliable gas guzzling V8s. Big AI is about coal rolling.
Too many of our business leaders have modeled themselves on the fantasies of a century ago, as titans astride a metal octopus, united by radio waves or whatnot, maybe as a Greek god. They were front and center on stock certificates in thr beaux artes style. They worship power, ideally highly centralized power. The is no place for something distributed like solar panels or locally intelligent systems.
We’ve seen this movie before.
P.S. Speaking of movies, I’ll recommend 1935’s The Transatlantic Tunnel for a taste of the times. It not only captures the dynamic spirit, but it’s also a women’s weepie.
Thank you for the summary.
I wonder at Microsoft’s approach. It seems better for Microsoft to concentrate on business apps and avoid consumer chat bots. They don’t do well in the consumer market.
Cloud service is nice to have. It makes the filing system more like PC’s. It is possible to duplicate files on Icloud. I don’t use upload or download anymore, just copy and paste from IPad to Icloud.
The default browser for this IPad 10 is Google, and Google Ai creates the Ai information. I can’t see why Apple would have to have their own, on top of what Google can find.
I think Apple is trying to price some of their hardware at lower levels. There is an Iphone at Walmart. This IPad was sold at Costco during Christmas season, mostly for gifts, to push people who don’t have experience with Apple into using Apple.
Apple is hoping people will buy services. I did get a cloud subscription shortly after I received this gift.
Apple spreadsheet Numbers works well, but it is a struggle sometimes to find out how to do things. It is kind of hard to work with the 5 column limit, but I got used to it. I used the drawing tool on some spreadsheets, which is a very nice feature. Not sure if the drawings will transfer to PC’s.
What five column limit? Numbers spreadsheets have any number of columns. Just drag the sheet horizontally. The interfaces is a bit clunky compared to VisiCalc on an Apple II but good for the new century.
That sounds like a five column limit for the column outline structure. Do they also have a five row limit? They probably consider five levels deep enough for an outline as each added column multiplies the columns in the displayed table.
Then again, I’ve never used pivot tables, but I did translate Apple II Visicalc to run on a Z80 Trash 80.
I saw that 5 column limit when reading about pivot tables. I wouldn’t try 1000 columns and 1 million rows without other information at about what file size limit I should use with this small IPad.
Five columns is about all I can see when there are larger column widths. If I wanted to use the data for a pivot table, it would not be necessary to re-do the table.
Where I saw this 5 column limit was in reading about pivot tables.
Here is a copy of an Ai message. My search bar: “is there a 5 column limit for a pivot table in Numbers IPad 10.
Response:
“According to users in the Apple Support Communities, Numbers on the iPad has a limit of 5 fields in the "Columns" and "Rows" boxes in the Pivot Options for Pivot Tables.”
I started using small computers when they first came out, early 1980’s. Visicalc was the first spreadsheet program, for the first IBM machine, predating Lotus 123 and Excel. I wonder if they just decided to use the same name, what a surprise.
Earnings look great for the Magnificent 6 (Tesla is a train wreck in an earthquake). But depreciation spreads out cap-ex for years to come; it will be a weight on future earnings if future AI revenue doesn't come. Also, the OBB Bill accelerates cap-ex tax expenses, so lower taxes boosts current earnings too. Everyone else's depreciation lagged expenses shows up as today's revenue for NVIDIA.
We can see it in the yawning gap between some earnings and free cash flows. AAPL and MSFT are exceptions in the most recent quarter ($ billions). Why shouldn't Apple rent someone else's AI, or let the user choose similar to search engines? They book the cash, others pay the expenses.
AAPL Net Income: $23B; Free Cash Flow: $24B
MSFT NI: 27; FCF: 26
NVDIA NI: 19; FCF: 26 (prior quarter balances this out)
Meta NI: 18; FCF: 9
GOOGL NI: 28: FCF: 5
AMZN NI: 18: FCF: 0
TSLA NI: 1: FCF: 0
A key point is that Apple is focusing on narrow AI for particular tasks. This echoes the GOFAI era of Roger Shanks' "scripts". It is far easier to reduce ambiguity when you know how to act and respond. Porting this approach to a private edge device makes a lot of sense, and is akin to my earlier comments that AI needs to be more niche and less general to be profitable.
Secondly, should there be an AI bust, those NVIDIA chips will sell at auction for cents on teh dollar. Apple can then snap up all it wants for its more generalized cloud AI.
But the critique of not delivering useful AI stands. Apple has failed to deliver production versions of its AI on devices to reliably achieve those narrow tasks. That is a major failure that needs to be fixed.
A friendly comment. I read you EFD and admire your scope and clarity. This posting although thoughtful, cogent, and as far as I can see (as a non apple person) correct, is (with all due respect) repetitive. Especially in your description of the cost of the arms race, and the Nvidia rent. I suspect you might be whelmed by the conflagration of Benjamins.
Just a suggestion.
Thank you❣️I am starting to look at a new computer, I’ve always had apple products. Due to the current yet ridiculous word police having privacy is very important as well. I encourage people to be regular listeners to your site or better yet support it! as I do. thanks again.
Very well explained and appreciated. I definitely agree that people will be focused on their phones for most of what they will use AI for in terms of non-business-related activities and likely to use their PCs or Macs for business planning, etc.
Apple started offering voicemail transcripts two or three years ago. Since I can’t usually understand recorded speech, this was immensely useful. It’s quicker, too. Apple’s Magnifier app has a mode to speak or display what the phone camera sees. They advertised it as an aid for the visually impaired, but it is hard to miss an entire business area. In actually providing useful AI, Apple does fine. Anyone who relies on a big LLM to summarize things is setting for a fall.
The AI push is essentially about undoing the PC rebellion and moving back to the 20th century model of centralized control. As the Chinese and Apple have shown, it is possible to provide useful AI at the periphery. The Chinese LLM performed AI at a moderately high level but with extremely low power consumption. Think VW Beetle in 1963. Apple is performing AI on wrist watches. Meanwhile Detroit is offering bigger, more powerful, unreliable gas guzzling V8s. Big AI is about coal rolling.
Too many of our business leaders have modeled themselves on the fantasies of a century ago, as titans astride a metal octopus, united by radio waves or whatnot, maybe as a Greek god. They were front and center on stock certificates in thr beaux artes style. They worship power, ideally highly centralized power. The is no place for something distributed like solar panels or locally intelligent systems.
We’ve seen this movie before.
P.S. Speaking of movies, I’ll recommend 1935’s The Transatlantic Tunnel for a taste of the times. It not only captures the dynamic spirit, but it’s also a women’s weepie.
You should notice that I don’t talk about the business case, because there is none. AI is about finance, not business.