> We in the US NEED to be able to manufacture for our own needs.

Really? Why?

The global system has worked really well for the last 70 years. Covid was a blip to supply chains, having them shorter may have helped get through the shocks faster but it wouldn't have eliminated the problem.

A core issue with being trade isolationist is that the US doesn't have the raw materials it requires to stand alone. No country does, event big ones like the USA. Some imports are always going to be required.

The second issue is why would you do it the expensive way. The USA has bauxite deposits and bauxite mines, but they are small and don't produce much. I'm sure with effort they could increase their output but it would be expensive. It is much much cheaper to buy bauxite from Guinea or Australia. Why would the United States choose to produce their own aluminum from their own bauxite? Why pay considerably more for the same commodity rather than just buy it? And of course it isn't just bauxite.

When PCs were introduced to businesses in the 80s and 90s we didn't see productivity improvements, on an economy wide level.

It seems super weird and backwards, hence it's referred to as the "productivity paradox".

Worse, the introduction of the internet to most businesses in the 2000s actually corresponds with a productivity slowdown, where it increased at a lower level than normal.

As the introduction of PCs didn't trigger economic growth and the introduction of the internet retarded economic growth it is ambitious to suggest that AI will buck the trend.

I know this seems super weird, I certainly feel sure we are all more productive with PCs and the internet, but the data doesn't support this.

The community said so. They frequently say stuff like "could place at risk lives of countless innocent individuals", but when asked directly how many people were actually hurt they say "no one". Remember secrets aren't absolute, government agencies are more than happy to release information when it furthers their interests.

* Pentagon spokesman Geoff Morrell "We have yet to see any harm come to anyone in Afghanistan that we can directly tie to exposure in the Wikileaks documents."
* Brigadier general Robert Carr, who headed the review on the impact of WikiLeaks disclosures, to court "I don't have a specific example,"

Carr's evidence in the Manning trial is particularly interesting, because he initially claimed that there was one Afghan who was killed due to the Wikileaks publications. However on cross examination he admitted that the individual wasn't actually named in the leaks. The Taliban claimed they killed someone named in the leaks who actually wasn't actually there, a claim which Carr repeated. it seems like both the Taliban and US governments had an interest in claiming related deaths to the extent of being willing to make one up.

Carr's evidence shows that the US government wants to link the leaks to deaths, is willing to provide information to link the leaks to deaths, even to the extent of claiming links when there actually weren't any. Their clear desire to publicize such a link and the fact that they haven't done so is strong evidence that no link exists.

The evidence around the 2021 deaths and the link to Trump is less clear, the primary source of information is a CIA cable that was partly reprinted by the New York Times. The NYT reported that "dozens" of informants were lost (died and double crossed), the cable had a specific number, it also recommended tightening a number of operational practices to improve matters. Other reporting has drawn links between those losses and Trump's actions. We may get more concrete information when the classifieds document trial eventually proceeds.

This is a weird debate to try to turn into a pro-Trump political argument.

There have been zero deaths directly linked to Assange's leaks. When asked the Pentagon said "We have yet to see any harm come to anyone in Afghanistan that we can directly tie to exposure in the Wikileaks documents." And given the efforts to charge Assange with a variety of crimes I believe if there was any links it would have been publicised. https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.bbc.com%2Fnews%2Fworld...

On the other hand a large number of CIA informants were killed in 2021. This has frequently been linked to Trump, conversations he had alone with Putin, documents in his possession and documents he published. https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.foxnews.com%2Fmedia%2F... https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.theguardian.com%2Fus...

Also the radio device they use is a module, probably a Broadcom module.

The protocol stuff is all handled within the module. To certify the device properly handles the protocol they don't have to redo all the testing, they just need to attach the existing Broadcom certification.

The FCC certification including Thread support likely means that the module they are using happens to include Thread support rather than any indication of Apple's support or intentions.

The summary talks about the 1:40 false positive rate, suggesting that it's a reason to stop. While not negligible 1:40 actually seems really good.

I would expect police officers approaching people looking for suspect X to be wrong most of the time, a 3% incorrect rate is a huge improvement.

Race factors are always a concern in systems like this, but it's probably being monitored and addressed. It's also replacing a human system that is known to have substantial racial biases.

That's why most of these suits includes samples of direct reproduction. The algorithms overmatch, so with the appropriate triggers the large slabs of exact source data will be produced. The courts haven't decided yet, but one argument being put forward is that the overmatches show infringing reproduction, not just learning.

The maths on using batteries to time shift supply doesn't add up, at least where I live.

Time shifting load on the other hand is far more worthwhile. Having massive consumers such as aluminum smelters running during peak power periods is hugely wasteful, idling them during those short time windows would significantly cut operating costs and significantly reduce load on the grid.

Or is it the opposite?

If the staff aren't in the SF office why pay a premium to employ SF staff?

Even airbags probably don't require an RTOS, you just want to fire them as soon as possible, and a collision takes a long time in compute terms. An RTOS is designed for consistent timing, not fast timing, and the use cases are actually fairly narrow and not what most people expect.

I have an engineering degree and an MBA (and an IT degree). Before I did the MBA I had a lot of vague bad feelings about MBAs and the folk they produce. Now I have very specific bad feelings.

I believe degrees shape you as a person. Engineers are trained, through classes, group projects, and a consistent problem approach to see the world in a particular way. It has a lot of significant advantages, engineers work well together because of that shared worldview. Legal degrees shape people to view the world and approach problems in a particular way, not too dissimilar to engineers. Other degrees such IT/Software Engineering influence you but don't shape you as much.

MBA degrees shape people. The consistent theme through an MBA is optimization, optimizing a company to achieve profit. They talk about there being other things that are important, but then the assignment is to make you optimize X for profit. It appealed to my Engineer brain, how to create a system that achieves an optimal outcome. The basic knowledge they give you on accounting, finance, economics, statistics, operations, strategy, leadership, negotiation, hr etc. are all tools for your toolkit in crafting these systems, and the exercises are almost always focused on using those tools to maximise profit.

So when a person with MBA training sees something, a system, a rule change, an emerging problem, anything... The training kicks in, how does this thing impact maximising profit, using the perspectives and tools available, how do I adjust my systems?

The process shapes someone, it doesn't rebuild them, so their underlying beliefs and drives are still intact. The training can also easily be reoriented to optimise other systems. But an MBA also self selects for people who are more likely to be profit focussed, it is a business degree, and then it groups them together to influence each other.

MBAs are trained to optimise capitalism, exploit its flaws, and extract maximum value. I don't think that is intrinsically evil, but it certainly isn't intrinsically good.

Another issue which is a significant problem for consumers is the wildly clashing life spans of smart devices.

Smart device interfaces are essentially tablets, similar hardware and technology. Most tablets aren't expected to last more than three years, developers issue a new product each year. Issues such as a lack of security updates and software incompatibilities rapidly increase after three years.

Fridges typically last about ten years, older ones are common but generally a replacement is suggested at the ten year mark.

Users are thus stuck. Either they cycle their smart fridge every three years, significantly increasing the cost. Or they let the smart aspect atrophy, after ten years the smart portion is almost certainly non-functional and they just have an ugly and expensive fridge. Neither solution leads to happy consumers.

Fair, but Microsoft actually has a good idea of the real identity of people committing to Microsoft code. Various forms of ID, background checks etc. Jia Tan (JiaT75) is just a name with no background, nobody seems to know who they are, what they look like, where they live, or who pays their salary.

Debian requires verified real identities to become a developer, but I'm not aware of other non-commercial projects that do. Maybe identity verification should become the norm.

I don't think the sharing customer information is necessarily required.

For a simple algorithm: inputs -> algorithm -> outputs

If everyone is using the same algorithm, and the algorithm is fed the same set of inputs then everyone will get the same output.

That has the practical effect of aligning everyone's price and removing real competition from the market.

The legalities seem far messier. Directly cooperating to set prices is clearly collusion. The DOJ seems to be arguing that cooperating to use the same algorithm in order to get the same prices is also collusion, which seems reasonable. The messy part is the cooperating, was Rainmaker deliberately chosen because they knew the other hotels used it and adopting it would lead to this outcome? And is that illegal cooperating, or legal mimicking? The smoking gun would be an agreement between the hotels to all adopt the algorithm for this specific reason, which would be fascinating to see.

This evil cloud exploitation argument rests on two assumptions, which aren't true in Redis' case.

1. That the company created the software and thus has a moral right to profit from it. That's not true in this case, Redis was created by Salvatore Sanfilippo. The company Garantia Data provided a cloud hosted version of Redis. Garantia Data then employed Salvatore and changed their name to Redis Labs, apparently Salvatore objected to them changing the company name to Redis. Salvatore then left the company, leaving stewardship of the Redis project to the company, and Redis Labs renamed themselves to Redis. The Redis company have unquestionably invested significantly in the Redis code base, but they aren't the initial creators or the sole contributors, there are multiple external members of the core development team including staff from Alibaba and Amazon.

2. That Amazon is distributing Redis for a profit but not contributing back to the community, freeloading off the work of others. That is demonstrably not true here, Amazon has staff working full time on Redis, they have contributed significant features, Amazon staff are on the Redis core development team (or were, until the license change). As far as I can see Amazon has been a positive contributing member of the Redis community, doing exactly what many people have previously argued they should be doing.

I'm not trying to pretend that AWS has been perfect here, but they also don't seem to be the obvious bad guys either. What they are doing looks the same as what Garantia Data did, providing a cloud hosted version of the product that they didn't create but do significantly contribute back to. Redis (company) also has a history of behaving badly, they introduced Redis Modules with a non-opensource licence, then spent years complaining that AWS Elasticache was not "feature complete" or compatible because they didn't support the Modules that had been explicitly licenced so that AWS couldn't use them.

I have high hopes for the Valkey fork by the Linux Foundation, the product direction will probably be improved out of the control of the Redis company. Hopefully in the future they can work with the contributions from the cloud hosting company formally known as Garantia Data.