compounding_it

Yes AppleTV+ is actually a major financial problem for Apple.

The strategy Netflix and Prime applied was that they just loaded it with average-costing in-house content and then a few were hits. With Apple they seem to be spending a lot on individual shows and their library is still small compared to others. I guess they can't really do anything average according to their philosophy. This is unfortunately not working in their favor with TV+.

Old chips cost Apple more money now not less. The A15 and 64GB NAND probably have low volumes and cost Apple more money going forward. So apple likely just upgraded to what the supply chain can reliably do for a while. Prices might go up even in six months to a year. So instead of buying a 200$ bumped to 250$ A15 Apple TV in six months they are probably upgrading it to make sure future price bumps remain relatively low and justifiable for upgraded hardware.

Very interested to know what TV manufacturers are doing who have heavily advertised their Android UIs which are bread and butter of collecting data. TV panels themselves have no reason to have price increases in memory shortage.

apple probably was losing money on the old Apple TV with old 4gb ram and 64gb storage in low volumes anyway and they probably would have to pay even more going forward. the upgrade seems to make sense from supply chain standpoint to standardise storage and chips across devices.

If anything apple is probably bring value to new buyers of Apple TV in this market. At 299$ you are getting hardware that will last quite some time as opposed to a 200$ device that might cost maybe 250$ in six months and give you lot less hardware capabilities.

My guess is that memory is the real reason for the price upgrade. Apple probably added some things to mask the price increase but the real reason is the chip. The old 4GB ram in A15 Bionic probably bumped in price so it went up to 199$ and now the new chip costs even more. Also its easier for Apple to make the same chip in a lot of hardware as opposed to having multiple chips being made, especially when its fused with memory and thats costing a lot.

>Where do they get that pricing from?

I cant buy the LG doorbell camera at my local mall in a fancy store. I had not even heard of it until now.

Volumes of 100s of millions on chips and volumes in hundreds of thousands will have a huge price difference. We are talking 2 orders of magnitude higher.

>AI to backdoor designs

correction: AI to backdoor vibe-coded designs

if anyone has read the trusting trust paper, one will know you can theoretically create code where vulnerabilities are hidden in plain sight. But that would mean its designed and reviewed by people who don't understand it in the first place.

The arrows on newer keyboards are gone. Now it's just text. I wonder if that changes anything because Enter doesn't really sit well with the arrow next to it as entering data need not necessarily cause a carriage return.

Same here. Have a Base M1 Air. Will trade that for M7 Base Air maybe if I can drag this one for a while. I can't go back to computers with noisy fans. It seems too last generation since the M1 launched.

Is it safe to assume at this point that GNU and the kernel accept AI code but there is quality control compared to private companies ? What I am hearing (and what friends talk about) is that with AI everyone is hooked on quantity while quality has taken a back seat. I would assume that the quantity and features don't bother free software much so they can take their own sweet time with these things.

From a psychological standpoint, maybe customers thinking 26 is old on their brand new device in 3 months isn't exactly good. People might start expecting 27 in 3 months when actually 27 is going to be around for most of the year.

More customers is generally a good problem to have in most businesses. Just that the situation is very paradoxical given the supply shortages.

Because the verification would also be done by something non deterministic and then that’s a paradox.

>optimization is by far the easiest part of the process

Not if the whole thing is architected poorly but was a requirement of the hour so it became big. Then optimisation becomes an art, but definitely not the ‘easiest part of the process’

> Hugging Face turned down a $500 million investment offer from Nvidia late last year that would have valued it at $7 billion, the Financial Times previously reported. Hugging Face said at the time it did not want a dominant investor that could sway decisions.

So instead they sold themselves to the same investor completely ?

>local models run on your mac for your workflow.

10 years ago 32GB ram laptops sounded too much. 8 was enough. These days even I would get that much ram since it’s soldered. 64GB is higher end.

In a few years we should see such high end hardware commonplace. Working with a local LLM to get work done is the ideal way to go which has mostly hardware limitation as of now that gets solved in due time.