It's not quite that high where I am but it's orders of magnitude more than 1/100, probably 1/10, but growing fast. I was at the Tesla center yesterday and they delivered 90 cars and were scheduled to deliver over 100 today. Considering the average dealership moves about 100 cars/month, that Tesla center is doing 2 months of business in 2 days. Granted the tech said it's not usually that high, but it gives an idea of the velocity of the transition over to EVs, at least for new car buyers.

I know here in America industrial rates, not residential, do factory in amperage or kVA. They still charge for consumption, but the rate is determined by the peak amps they draw in a billing cycle. At least that's the way it was 20 years ago. It goes both ways too, where if the amp draw is too low, penalty rates kick in. I'm surprised they apply this to households in Europe, there's a tiny difference between 100A service at a home and 4000A service at a factory.

Sure, but that's 240V@16A, that's 2x the juice that the standard North American plug which (comparatively) dribbles out 120V@15A. 240V@16A is more than enough to recharge the majority of EVs overnight with the exception of monster 200kWh battery EV trucks like the Lightning, Ram or Hummer. At 240V@16A, we're looking at 3.9kW, so assuming 8 hours, based on current level 2 chargers, one could re-add 120 miles / 200 km. Given the majority of commuters in Europe travel 40 km / day, it's likely most could easily live with charging off a standard 240V@16A socket.

Current assembly lines run tight schedules; shipping components adds significant lead time and risk. Another option is to automate the production lines extensively, which targets the labor cost but doesn't fix the component supply issue. The Chips Act may help, but for now, lead time from suppliers is going to be a challenge and a risk.

There's one big problem with Mexico though, it has nowhere near the population of China and India. It's unclear if Mexico even has large enough population centers (outside of Mexico City) to staff a Foxconn factory. Mexico has a single mega-city and 12 mid-sized cities with populations between 1-2 million. Compared to China and India, these mid-sized cities may simply not have enough workers. China has over 50 cities with populations over 2 million. Shenzhen (Foxconn HQ) has a population of 17 million, which is 90% larger than Mexico City. Trying to relocate a single assembly line from an area with 17 million people to an area with 2 million is going to create staffing challenges. To give a high-level view, Mexico's entire labor force is around 59 million, with a labor participation rate around 70%, so that means 42 million total available workers. Foxconn has approximately 800,000 employees. So if everything moved to Mexico, Apple would (indirectly) employ 2% of Mexico's labor force. Mexico's unemployment rate is currently quite low, so even hiring those employees away from their current employers could be problematic from a cost-advantage perspective.

Insanely high levels of automation (like Musk is trying to achieve within Tesla) are probably the best bet for moving manufacturing out of China and India. If we can get to high-dexterity polyvalent robotics, it's not impossible that these assembly lines could be onshored, but I'm guessing we're still 8-10 years away from assembly robots that can handle miniature component assembly with enough dexterity and complexity to be a viable alternative to human assemblers. Ironically, Apple's disassembly robots could be precursors to future assembly robots, however taking something apart is orders of magnitude easier than assembling something.

I was thinking the same, why put this in Albany? No offense to Albany, been there, it has beautiful lakes and hills, and people are friendly, but it's not on any top move-to destinations list in the US. I could say the same about the fabs going into Ohio. If one is building a facility where there is no existing local talent pool, then at a minimum pick a location where people are actively trying to relocate to. I've never heard anyone in tech say "I wish I could to relocate to Albany or Columbus". I get the allure of tax rebates and government kickbacks, but if there's anything to learn from the "it's the next Silicon Valley" hype of the last 20 years, it's that building it does not mean they will come. The smart money builds where people already are or where they're going, building where people aren't and aren't moving to is not a path to success.

"Not sure how crappy they are, but they must be cheaper, as almost all big US companies have IT and development operations outsource/offshored."

Fixed that for you.

EVs have ~20 moving parts, ICEs have ~2000. EVs have no transmission, gears, oil changes, timing belts, head gaskets, etc. The list of things that can go sideways in an ICE vehicle are almost endless. In an EV, it's either the battery, the computer or one of the motors. My experience is that EVs are effectively maintenance free. I've spent over $4K in service and out of warranty repairs for our ICE SUV this year, and that's a Lexus, top of the pile for reliability. Meanwhile I've spent $0 on our Tesla. I understand the pushback against EVs, ICE maintenance and repairs are a reliable revenue stream for dealers and service centers. EVs will only replace a small fraction of that revenue stream. Those with a vested interest in ICE service and repairs have a right to be fearful and angry, but pushing the false narrative that EVs are more complex than ICE vehicles, that's.a blatant falsehood.

I couldn't find out from the linked articles if they're using modulation or pulsed. I have to assume it's pulsed since space "detritus" would impact modulation more than pulses. The other thing that's puzzling is how diffraction is handled, the beam should be insanely wide at 10 million miles, even with special focusing lenses. Unless that's a feature, so you don't need to hit the bullseye while hurtling through space, but then that's hogging bandwidth from anything else in the same direction from the receiver. In any event, glad to see NASA is pushing the envelope with this, could lead to interesting stuff. The other benefit is that since they're in the photon game with this, they're also in the parking lot of the ballpark of quantum communication.

This has pork written all over it. A list of tech hubs that don't include a single tech hub, with the possible exception of Massachusetts and Miami if you squint hard enough. No, this is money that will simply get siphoned off into someone's pockets.

First off, $40M buys absolutely nothing, maybe a mid-sized office building or 20 top-tier engineers with materials and supporting staff for a year.

If Silicon Valley has shown anything it's that concentration of skills wins. Want to create a hub? Invest BIG in a specific area, $600M-$1B sounds ballpark, and ensure a core group of world class experts are in that area to draw from. Then you just have to hope the experts agree to go, that some kind of synergy develops, and that the timing is just right.

You can't just anoint North Haverbrook a tech hub because they're getting a monorail. This is just burning tax payer money to back pat some politicians.

The problem isn't demand, it's that Ford will happily sell you a fully loaded 4x4 V8 F150 Lariat for $60K. However, for a similarly spaced out Lightning Lariat the want $80K. That's.a 33% premium, which paired with a 7% financing offer, leaves only those comfortable with a $1425/month car payment to buy their EV truck.

The GM Silverado EV is even more whacko with a BASE price of $80K.

The addressable market for EV trucks at >$75K is tiny. Another problem with hyper-expensive EV versions of gas trucks with $35K base models is that people spending >$75K on a truck don't want to be lopped in the same Venn diagram as a $35K truck.

Tesla's best selling model is a sub $50K all wheel drive SUV with 300 miles of range that doesn't even have cloth seats as an option. The cheapest Ford option with all wheel drive and a 300 mile range is $70K, and you can only get cloth seats at that price.

So it's all whiney nonsense, Ford and GM don't WANT to sell a $50K EV truck people want. They're in the "nobody wants these" phase, while ignoring the "for a 33% premium over our gas models". There's a non-zero chance they will get hollowed out by the Cybertruck *if* Tesla manages to deliver a $50K 4x4 with 300 miles of range that (gasp) has seats that aren't cloth.

NoSQL was a terrible idea from the start. That said, memcache, the original KVS still has its use cases. However, trying to bolt query semantics onto a KVS and claiming "it's not SQL, it's an API" was always a fools errand. I can see untangling metic peta-tons of "unstructured" data into normalized SQL schemas becoming big business as companies come to see the limitations of NoSQL.

You can see that today, ask ChatGPT "Write me an app in SwiftUI that can retrieve and store public stock trades in order to generate buy and sell recommendations based on momentum.". The output is certainly "an app", but it's not even a basic MVP, it doesn't retrieve, store, analyze, or generate anything. Like you pointed out, "coding" isn't the hard part, it's the architecture, design, filling the missing gaps, and polish that's the hard part. Currently, AI does an OK job of creating an absolute bare bones starting point, but it's not much more helpful that it creating a book with chapters numbered 1-10 and a bunch of blank pages.

"no VM anything"

The whole Docker running a VM on a VM has always bugged me. Docker is still a VM, but at least on dedicated hardware it's only one abstraction. Everything comes in cycles in IT, I imagine we'll see some form of reversion to hosted hardware, possibly paired with VMs being replaced with innovation in process sandboxing.

Exactly, the M1 Pro came out in November 2020, that's 2 1/4 years ago. The M2 is 35% more powerful than the M1 at graphics tasks, and it's been shipping for 8 months already. So kudos to AMD for pushing the envelope, but it's most likely that this Ryzen chip is at best comparable to the M2. However, that implies that it will be significantly slower than the rumored M3.

I wasn't there when Wall made his design decisions, but I'm guessing it went like "whatever it takes to make a function fit in 80 characters on my VT". That said, it's a pretty safe bet you'll find it on pretty much every piece of non-Windows hardware in existence, so it's a good language to be familiar with, even though deciphering someone else's Perl is in no way guaranteed.