If they would reprocess it, much of that 3.6 million pounds would become either valuable nuclear fuel or totally inert scrap.

The rest would only last 200-500 years before becoming totally inert.

Who wizzed in your wheaties?

And I stand by what I said. Practically no phone that still works at all does it at the low level of a capacitor anymore.

Exactly. For example, spreadsheets were often used day to day but the accountants kept the actual books as books and had the final word. That changed over time, only after spreadsheets proved out (and got the bugs out).

Also, humans stayed in the loop. Sure there were occasional funnies like a medical database claiming a few male patients had a history of pregnancy, but humans were in the loop and cited those as obvious errors. There were still backup paper records to fall back on.

But a lot of people when they hear someone say "it's nowhere near ready for prime time" will claim the person is saying "It'll never work".

Back in the day when the "Board of Education" was an actual wooden board.

If they really wanted to serve the user, it would understand "Hey Google, lockdown now" and would power off and require a PIN to restart.

Long press doesn't do anything so low level anymore, it just signals the management processor in the radio that it should force the application processor off. The management processor keeps running.

The power button on the phone really is just a suggestion. To actually power off the whole phone you have to remove the battery (if possible) or let it run down to zero.

This is what happens when you give an AI agency.

You've got questions, we've got cellphones.

In a real sense, the back end recognition that Flock does is exactly stalking on a massive scale.

It's not a matter of not being seen or even not being photographed every now and then. That level of privacy is not available in public.

But even in public, until Flock and it's ilk came along, we had the privacy of being just a face in the crowd. The people who saw you on main street were different from the people that saw you on 2nd street, were different from the people who saw you leaving your home. There was no database and no recognition software to stitch all of the sightings together to form a coherent record of your movements.

It's that tracking and recording that people object to.

The compute cost for DeepSeek V3 was $5.6M. Training costs for traditional US frontier labs model is about $70 to $200M+, with estimates of up to $1B+ to train the next generation models. It's expensive, but it's not even in the ballpark of "There is no viable economic model to do so." Google revenue in Q2 was $1.32 billion *per day*. $1B is trivially less than a day; $5.6M is just 6 minutes of revenue.

The cost problem is in the compute buildout and the subsidized compute to capture users.

> It remains to be seen if even using their models can be profitable,

I presume we can stipulate there exists some tasks that are cheaper to do using an LLM than doing it "by hand". As a such, using an LLM model can be profitable. To me, it is also obvious that the business of serving an LLM model can be profitable. There's large benefits to being able to get tokens quickly, and you can expect people to pay for that, even if we presume they could run the model locally with no maintenance overhead. There's even cases where using the cloud is cheaper than you can serve locally at all - I saw one guy that measured his added cost of electricity when his computer ran a specific LLM, and it was more expensive to run it locally than to buy tokens from the cheapest inference provider for the same model, presumably because they had better hardware and cheaper electricity.

This. Madoff didn't get in trouble until he started fleecing people who were richer than he was. Then it became an issue.

Had he done those things, you would have heard me complain.

I simply don't remember being hassled to show anyone my vaxination card.