This strikes me as an attribution problem e.g. Within large companies business units are treated as standalone financial entities so you are either making money or a cost center. I think what could be happening here is that the kubernetes group is generating what look like un-paid-for costs, instead of being back-credited for ALL of the compute they are selling via the worker VMs. I would argue that people come to GCP for GKE which means it should be seen as a revenue generator, in that light it makes sense to hold the loss leader of the control plane open and 'free' to grow GKE as quickly as possible. This decision is bad for business.
That seems reasonable, but easier still is putting your thumb on the Gigafactory 3 from Tesla. As long as Tesla understands that things will go smoothly as long as Starlink happens to not work there, it won't.
Curious about the 'doesn't work while moving' that seems to me to be one of the advantages of a phased array. Since it is software derived aiming I would guess it would excel at adjusting for movement. Do you have details on that I can read on?
I think it might be time to look again? Cost of panels is extraordinarily low and cost of installation (e.g humans) is not on a similar curve. I looked at price/watt and went ahead and installed. The payoff period is ~11 years, 1 1/2 years in. Warranted to 20. You could take panel price to zero and it wouldn't have changed the math much.
I find that people like to think of this as a theoretical problem, when it is tractable to solve the bulk of it for a lot of folks. Thinking of the average price of a US vehicle $37k, here is what we did with that same amount:
Installed 9.4 kw of solar panels (32 panels) on our roof which generates the same amount as all of our electricity usage on an annualized basis for $24k. Bought a used Nissan leaf for $10k. Refuel at home, from the sky. No oil changes. Prepaid power for life.
I know a lot of people don't live in such a sunny state as Arizona but given that I live in Oregon they can take heart.
Then I got 11k back from state and federal subsidies.
I wish electricity and gas prices would go up, that's the only way to drive additional investment into a non stupid energy future.
So your argument is that because cars have upsides, the downsides can't be greater? The downsides of making cars central are huge, and because of that large progressive cities are removing them. That doesn't sound like hyperbole, that sounds like the pendulum swinging the other way because on balance, they are negative.
You could buy a used leaf today, it has USA safety features, is at that price point (mine cost 9.9k after fees, you can find them now for less) and has more than 100kms range. The leaf is better than this car in a hundred ways.
I think the parent was implying that the show's effect was time shifting imminent suicides to 'now'. Since the consumption of streaming media is big-spike-on-release + long tail, that big spike when the show is released would still be visible. It would look like a spike and then dip for suicides averaging to the trend line on a larger time scale.
That's where we disagree. You think power storage will take a magic physics solution, and I think it is normal engineering (pumped storage, rail/gravity potential generators, utility scale battery banks for leveling).
I'd go one more explicit step: the carbon we emitted made us wealthy while externalizing the costs. The recipe for gaining wealth at present for poor countries is, emit a shit ton of carbon in the pursuit of growth. If we're serious about reducing emissions we'll have to provide a different recipe.
That's right. All my neighbors use the excess solar power during the summer, and during the winter I use more grid power than I create. I get credits in the summer and use them up during the winter.
A giant leap in solar efficiency doesn't solve the storage problem, and the storage problem doesn't need a magic physics solution, just large scale heads down engineering. My point is: solar is efficient enough now, and a sound economic choice. I make enough power in a year just with my roof for my house and car. I do need a storage solution to do power leveling, which is what I use the grid for now.
No one who charges an EV away from home is charging at that rate, that's for an at home 'trickle' charger. Look up any L2 charger and you're well above that rate, and rates are only going for public chargers.
Why is it not feasible? I run 32 panels on my roof, in the pacific northwest which is known for its rain, it covers 100% of my residential and electric car usage when considered on an annual basis. If roofs required solar panels we would have a large part of power generation done at a much lower intallation cost and could focus on large scale energy storage projects instead. My payoff period is 11-12 years, system is warranted to 20. No breakthrough required.
This just seems like common sense though? There isn't a conflict here that I can see. Use renewables to eliminate some carbon, add natural gas plants (which are cheap and quick to build by comparison) for leveling. Dump coal. Then build nuclear which take a decade or more to replace natural gas plants.
LEED is like an energy star rating for a home. Mine got a silver I think (low 50's). So it could be mandated, display the houses result alongside the listing perhaps.
But the article doesn't say "look how tough we have it" it says "look how broken we all are." And not many people want to look, instead we want to do the endless phone upgrade cycle except with our lives.
No, because this isn't a binary result. If group a and b both have bias 1, that bias won't be cancelled. But if they differ on bias 2, the combined group has less bias than a or b while not having 'cancelled all biases perfectly'. Perfection is a false goal while 'lower risk' is attainable and worthwhile.
And the forest fires and beetle vulnerability are not unrelated to current climate change (at least in California last I read), so that might help close the non-anthro-centricity argument they are using for different accounting methods.
Thanks for posting this, I have started working up some cloudformation scripts with the goal of getting spot instances into play for GPU nodes on AWS @ https://github.com/frankamp/fast-ai-aws-advanced not quite there yet, but using layering to get separation of concerns means a user can swap out for a bigger gpu temporarily without losing anything. Cheers.
This is a key point, and points to so many thing being broken at the same time it hurts my head. The real smoking gun would be to find some sort of defacto evidence that doctors got the hint that denial gets them more business. That would sit in the next layer up, the assignment of cases. Alternatively, could it just be a big pool e.g. the race to get through them is a cash grab, so the faster you go you'll get to drink more of the milkshake? In that case denial is more of a way to avoid scrutiny.
Point three of yours is the primary concern with 'superhumans', so its unlike vaccines entirely. It's not about the individual at all, its about the new species we'd be creating, that would reproduce on its own. How many edits would it take?
This reads like a snl skit, where they keep repeating "they did it themselves". Next scene, alien archeologists: "huh, each human in the house apparently caused this 10000-year-before-degrading-polymer to be produced for deadpans and looks at the camera a beverage!" Another team: "Ooooooh we found a lot of those in giant sea creatures carcasses!" Together: "They did it themselves!"
Classical critical thinking skills are great, and no match for reinforcement based suggestion engines. You might critically pick the best item from the 5 pieces of shit you are offered next, but comprehending how your viewing habits are being shaped over time by moving the _neutral_ point a half-a-degree is another thing entirely.
That's good, and here is some more including different ways to measure implicit bias that have emerged. The point stands that people aren't objective:
>White applicants get about 50 percent more call-backs than black applicants with the same resumes; college professors are 26 percent more likely to respond to a student’s email when it is signed by Brad rather than Lamar; and physicians recommend less pain medication for black patients than white patients with the same injury.