Cherry-picking a single packaging announcement from a single vendor does not disprove anything. Obviously there's trade-offs between battery metrics, and obviously commercial products lag behind lab results, but that says nothing about the industry's continual improvements over time - and neither do your quotes. Perhaps you should be reading them more closely yourself.

While new chemistries and packaging arrangements can certainly help, there have also been steady improvements in existing cell chemistries, and in particular the anode and cathode materials & structures. Some more citations for you to read:

RMI, 2024:

Ars Technica, Batteries keep improving

Trends in the Gravimetric and Volumetric Energy Densities of Lithium-ion Batteries Over the Past Decade, 2023

Commercial cells available with 400-500 Wh/kg densities, up from around 80 Wh/kg 30 years ago.

LTO batteries have great charge times and long cycle life, but poor energy density (worse than NiMh) and high cost. Supercapacitors are cheaper and even faster to charge, but have even lower energy density. EVs generally use NMC (3x the energy density of LTO) or LFP (150% density with much lower cost, albeit slower charging than NMC).

In actual fact, energy density of lithium-ion cells (not packs) has grown from 80 Wh/kg to 300 Wh/kg in the last 30 years, with cells reaching 700 Wh/kg in the lab.

May I suggest backing your opinions with citations in the future, as loud but unsupported assertions are generally ignored.

More than a decade I think - certainly if you're talking launch systems. The engine in your link has a thrust described in the single-digit Newtons, whereas we need megaNewtons to lift any significant payload to LEO. Fine for deep-space long-duration missions where efficiency is key (assuming we can find the Helium 3 to fuel it), but not suitable for launch.

The Saturn V was a huge achievement, especially for its time, and a credit to the engineers who flew it so successfully despite its many issues. It could lift an impressive 140 tons to LEO. But it was also ~$1.5 billion per launch in today's dollars, and that all got tossed in the sea every time.

Starship is much more capable, around twice as powerful, yet is a fraction the cost at ~$100M to build and launch. It can lift around 230 tons to LEO when fully expended, so it is more than capable right now of lofting the entire Apollo mission stack, lunar injection stage and all, along with a Lunar Gateway for good measure. Its Raptor engines are the only full-flow staged-combustion designs to ever fly, and have a power-to-weight ratio nearly 50% better than the Saturn V's F-1 engines, despite being 1/30 the cost to build.

But Starship's real value is in reuse - and this is why they keep doing test flights. It's designed to land back on its own launch tower, then get refueled and relaunched on the same day. When you can use your expensive ship more than once, then your marginal flight cost is not much more than the cheap propellant - and now your flights are $10M or less - for over 100 tons to LEO. Two orders of magnitude cost reduction changes the game completely, for so many things.

Then add in orbital refuelling, and now anything you can get to LEO, you can send anywhere in the whole system. Apollo put 15 tons on the Moon, while Starship could deliver 100 tons, a dozen times, and still be cheaper than a single Apollo mission.

Half-Life 1, yes. Half-Life 2 required Steam to activate it though, even physical copies included a Steam CD-Key.

Unless you liberated a copy while sailing the seven seas as a privateer of course. Which was somewhat more common before Steam hit its stride.

The Soviets supported any potential source of discord and conflict in their opponents, as do the Russians (and CIA) today. In this case it probably backfired on them, as increased environmental awareness (and concomitant environmental regulations) have overwhelmingly benefited the US. A 2015 OMB analysis of costs & benefits of Federal regulations showed a 250% to 1400% net benefit:

And more specifically, the Clean Air Act of 1990 cost around $65 billion over the next 30 years - but saved $2 trillion (not to mention preventing 220,000 premature deaths) - i.e. a benefit to cost ratio of 30 to 1 .

I fully expect they'll simply dismiss it all as "bogus" anyway. Some people just don't want to look too closely at their own beliefs.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fen.wikipedia.org%2Fwiki%2F...
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.imf.org%2Fen%2FTopics%2F...
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.iea.org%2Ftopics%2Ffos...
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.eea.europa.eu%2Fen%2Fa...
https://fossilfuelsubsidytrack...

TL;DR: Around $7 trillion in 2022 alone, globally. Consisting of both $1 trillion annually in direct subsidies, and another $6T in implicit subsidies (e.g. forgone taxes and medical & environmental costs borne by the public).

I'm saying that full self driving - as defined by the SAE, not the marketing term - does not exist outside of a car park, because the car requires a human driver. Tesla themselves state that the vehicle is not autonomous. The robotaxis promised years ago to earn money while you slept are nowhere to be found.

Maybe you feel that (as of 12.5) it could drive itself without a human in the driver seat - and for basic driving in common circumstances, I'd probably agree. A number of self-driving systems can do that much (see the self-driving Cannonball Run, currently led by a Prius with comma.ai's system with a 98.4% run). But the world is full of uncommon circumstances, and it has to cope with all those too, which has yet to be demonstrated . Other companies have received commercial licences for autonomous self-driving cars, and have completed tens of thousands of trips without a driver, but Tesla has yet to apply.

It's just not autonomous yet. Maybe "Actually Self Driving" will be announced in October, but I'm not holding my breath.

Well they kinda have done, the Falcon 9 is far and away the most successful rocket ever built, by pretty much any metric. It's also the only orbital rocket with any practical reuse (the shuttle did more but with hugely greater cost and time).

Starship is difficult to compare to other rockets, as it's unique in so many ways, from its assembly-line stainless steel construction, full-flow staged combustion engines, unique landing approach, and massive payloads, to its focus on rapid reuse of both first and second stages. Development time is longer than the Saturn rockets (if you don't count their Jupiter missile predecessors), but it's quicker than SLS (not counting its Constellation predecessor), for a vastly more ambitious design (and much lower costs). If and when they get that working, it will change the game again.

But Mars? Yeah, I don't think anybody's taking Musk's Mars timelines seriously :-)

Full self driving *Supervised* is there, but still requires human drivers to intervene semi-regularly, thus is classed as Level 2+, not even Level 3 limited autonomy. Still waiting for independent confirmation of actual safety, but Tesla hasn't allowed access to the data and Musk's claims aren't the most reliable.

The closest Tesla has to real self driving (with no human in the driver seat) is basically restricted to US car parks, and the human is still legally responsible for any faulty actions.

But that's exactly my point: it's not. The load curve is not the same, not anymore, and neither is the minimum line.

Yes there still IS a minimum line each day, so technically there still exists some kind of "baseâ load, but when that minimum keeps falling like this then it's meaningless. It's dropping so far right now with these sunny days and mild weather that we're curtailing like 25% of our renewables generation because the "base load" plants just can't get down to that level. The minimum demand is now getting so low, it's cheaper to throw away free power, because the coal plants will pay *us* just so they don't have to shut down.

You can't meaningfully plan around "base load" in that environment. The term has no further use..

Sure, but non-dispatchable plants are not load-following, and can be a problem when their relatively fixed supply exceeds current demand - too much power on the grid. France often has this issue with its nuclear fleet, which it solves by selling excess power to neighbouring countries for cheap.

And if that minimum demand for your base-load generator is actually negative? Or more precisely, if that minimum is below the fixed supply from your nuclear, coal, or other non-dispatchable plants? This is increasingly a problem my country's grid is facing - what used to be the base load has been pushed lower and lower every year. When it's that variable, it's meaningless.

The huge rise of rooftop solar here (now around 35-40% of households) is dramatically changing the market. Not only is this "hiding" demand from houses that cover most of their own daytime power needs, creating the duck curve - but they're all exporting their excess too, flooding the grid with super-cheap gigawatts that drive prices negative, which traditional base-load generation struggles to deal with. When renewable generation exceeds ~60% (we passed 90% yesterday), inflexible plants lose too much money to operate, and are being retired many years earlier than planned.

As a long time commenter at both sites, I'd have said that was the wrong way around. Politics is a much bigger factor on Slashdot, perhaps because it has a more diverse mix of left and right (a big reason I'm still here).

At Ars it's certainly true the audience is more left-leaning, but I would argue that they're generally much less interested in overtly-political discussion (there's enough of that elsewhere). You're a lot more likely to get upvoted for presenting facts there - if they're well-sourced and not blatant political stalking horses - whereas here people regularly get moderated Troll or Flamebait by one side or the other, regardless of accuracy or citations.