Yeah, I respect the hell out of them for doing the episode. Anyone who thinks old school Trek wasn’t silly as hell never watched much of old school Trek. It fully committed to being extremely silly and that’s great. I just felt like it could have been a better version of what it was.
And I stand by my comments about the Klingons. If anything, they should have been intervening to stop the Enterprise from closing the rift.
I gotta say, I really wanted to love the musical episode, but I actually felt like it was mid at best.
The songs weren’t memorable, and they mostly were just a character talking about their feelings. There was very little use of music to advance or explore the story. It was basically a standard episode with songs just kind of shoehorned in where they could. Compare with Once More With Feeling from Buffy, which I think demonstrates a much better understanding of musicals as a form.
Also, the idea that Klingons would be ashamed of singing is just insane. Klingons are the most dramatic culture in galactic history. Opera is canonically one of their favourite art forms. I refuse to believe that Klingons wouldn’t find the entire scenario rad as fuck.
Honestly, you can just skip straight to SNW, and you’ll thank yourself for it.
Discovery is a Star Trek show written by people who are embarrassed at the idea of being associated with Star Trek. Strange New Worlds is Star Trek show written by people who fucking love Star Trek.
“We fired our content moderation team because we believe in free speach… Wait, no, not like that!”
As others have suggested, OSMC is OK, but personally I prefer having Android so that I can use SmarttubeNext and access native apps for stuff like Jellyfin, Dropout, Nebula, etc. For years I played with various Linux options, but in the end I ditched it all for an Nvidia Shield and I couldn’t be happier with the results.
Do you also burst into astrophysics conferences and tell people that the sun is a burning ball of gas?
Please read what I wrote again, very carefully, and ask yourself “Am I actually contributing anything new to this person’s understanding of the topic?” Because everything you’re throwing out here is basically baby’s first blockchain, and none of it comes even remotely close to addressing my actual argument, which seems to have been a few steps beyond your grasp of the subject.
I actually have to disagree with the notion that blockchain (meaning distributed public ledger blockchain for these purposes) solves the trust problem. I’ve yet to see an example where this is actually true.
What public ledger blockchains do is move trust to somewhere off-chain. Consider a common example that’s used by advocates; tracking a banana from source to sale. The blockchain is supposed to create a completely untamperable record that proves the banana in your hands was ethically sourced. But in order for that to actually work, there has to be some way to prove that the banana in your hands corresponds to the record on the chain. And that proof can only come from human verification at each step of the process. So the trust is still there, it’s just in the humans verifying the accuracy of the records rather than in the records themselves. Which is basically how the current system works.
And you’ll find this same problem with basically any and every application of public ledger blockchains as a solution to problems of trust. In the vast majority of cases, sooner or later these trustless, decentralized systems will ultimately defer to trust in a central authority. How do you know your Bored Ape NFT is the real one, and not a copy minted by someone else on the same chain against the same image? You check it with BAYC through OpenSea or whatever platform they’re using. Ethereum has been forked multiple times. How do we know the current “Ethereum” is the true and real version of the ledger? Because the Ethereum Foundation says so.
These systems only solve trust in the same way that sweeping all your trash under the rug “solves” cleaning your apartment.
You’re right, that’s definitely what Sam is trying to do here. Unfortunately for him, he’s still an idiot, and he’s inadvertently telling on himself here by openly confirming what’s been well understood for a while; ChatGPT simply is not profitable to run because the models are so stupidly inefficient. That’s a real problem, and one that they’ve shown no meaningful plan for solving.
If AI cost peanuts to run, this would be a very reasonable point. But it doesn’t. It’s staggeringly expensive to operate something like ChatGPT.
So any use of genAI has to consider the question “Do the benefits provided actually justify the cost?”
Obviously, in a capitalist society this turns into “How can we monetize this?”, but even in a fully socialist society it would still be necessary to ask if this technology is actually providing sufficient societal benefit to actually justify the material resource cost of running it.
This is exactly the problem. There are plenty of people who will crawl out of the woodwork to tell you how they’ve found a way to make AI “useful”, but very, very few could put their hand on their heart and say that it was “essential” to their workflow or their own happiness and wellbeing in any meaningful way.
Which is fine in theory, but “expected” based on what?
They haven’t demonstrated any ability to meaningfully improve their models (“meaningfully” meaning "sufficient to actually address the very serious concerns about their practical usability), they haven’t shown any ability to meaningfully capture enterprise sales for their API, and their conversion rate on free users to paid users is abysmal. Their only stated plan to increase revenues is doubling their prices, which given their already terrible user retention doesn’t actually seem like a reliable way to bring revenue up. Jacking up prices only works when your users find you indespensible, and everything OpenAI offers can be found elsewhere for less.
I am sure you will find ways to argue against this, because when someone’s mind is made up there’s little space
There are literal pages upon pages here of people trying to convince you that you’re coming at this whole thing wrong. Not to mention those chiming in with their upvotes and downvotes (a misuse of that system, sure, but they’re raising their voice none-the-less). Are you really sure that you want to start throwing these particular stones from inside your glass house?
Your answer isn’t good enough either. Aren’t you forgetting application servers, web servers, load balancers, Cloudflare, firewalls and all that stuff which allow a database to just use 0.1J? Because if we are talking VISA and banking scale of transactions that’s what it takes.
I’m not “forgetting” those things, because they’re simply not relevant to what’s being discussed. A web server doesn’t “allow” a traditional database to use any more or less power. A web server is a web server. A firewall is a firewall. They’re not in any meaningful way connected to the transaction layer that we’re discussing. Blockchain validator nodes also sit behind firewalls, if the people running them know what they’re doing.
Besides, it’s just missing the point. Traditional databases are good and best at what they do - address traditional problems. Blockchains address different problems, so comparing them for completely different use cases won’t work. You can compare MySQL vs Oracle vs PostgreSQL that way.
Again, this is just a handwave.
If your argument is “Public ledger blockchains can be just as efficient as traditional databases”, which is the argument you previously presented, you need to actually demonstrate that.
If your argument is “It doesn’t matter if public ledger blockchains are less efficient, that inefficiency is worth it for the unique benefits they provide” then, first off, why did you make the other argument originally, and second, what have you done to actually show that only a public ledger blockchain can solve the problem you’re describing?
So, if I understand your pitch correctly (and, let’s be clear, this is information that needed to be presented right off the bat if you actually wanted to communicate this idea effectively), you’re envisaging a model where you sell some kind of hardware, presumably a complete solar panel kit of some sort, which then acts as a uniquely authorized validator node on your network, while also accounting for each unit of power pushed by that panel. As validator nodes, each panel contains a full copy of the database, and acts to verify new transactions, ensuring the integrity and security of your blockchain.
I’ll allow, for the sake of your argument, that your keys and code are sufficiently secure that you’ve accounted for basically any possible hacking risk. We don’t need to get into that argument. While in practice perfect security is impossible, for now we’ll say that your hypothetical security is “good enough.”
Right off the bat, we run into the following challenge:
Its remotely possible that the economics of the whole thing makes the latter option unappealing, but if so, I can’t see it. At best you’ve basically removed the incentive to use solar that the scheme is supposed to offer.
Another technical issue with this approach is that you want these devices to be usable wherever the sun shines, but in order for them to be able to each act as a validator node they have to each contain a full copy of the database, and that means having at least a decent internet connection if this system is ever supposed to scale. That isn’t going to work out at the cabin.
But even supposing those problems are solvable, and supposing that you can solve the problem of how the power gets from the panel to the charging stations without going through the local power company, we’re left with this question: Why blockchain?
You say that you want this to be distributed, public, not under the control of any one entity, but your keys would have to be authorized by a central authority. You would have to be the only producer of these devices to ensure that some unscrupulous individual doesn’t build a box that runs a hundred validators at once, exposing you to sibyl attacks again. You would also have the ability to revoke any key at any time. There would be nothing truly decentralized about this system.
I addressed all of this in this other reply: https://sh.itjust.works/comment/15924074
Basically, handwaving at “Ethereum is proof of stake now” just isn’t good enough. The difference in scale of power consumption between public ledger blockchains and traditional databases are so vast that even the most optimistic models for reducing their inefficiencies still only get you to “pretty bad” at best.
The most rosily optimistic estimates of proof of stake’s reduction in Ethereum’s energy costs (that I’ve seen) put it at a 2000x reduction. That means that in theory, if all of those gains were realised, and if we start with the numbers I previously cited, Ethereum might hit the same energy cost per transaction as a pretty inefficient traditional database setup.
Except that transaction rates are fixed in public ledger distributed blockchain systems (because every validator node has to have time to clock in with their results), so as the network scales up the cost per transaction also scales up. I’m actually doing Ethereum a favour by using old numbers there, because it’s final cost per transaction prior to the proof of stake switch was certainly much higher than it was at the time of that snapshot.
Traditional databases scale in a way where the economy of scale works for you rather than against you. The bigger you get, the lower your cost per transaction even as your total costs increase. Blockchains anti-scale; the cost per transaction goes up as the network gets bigger.
Well done for being a victim of their bullshit.
Mercedes says “ten years” instead of “two” for exactly this reason. Compared to the “loudmouth Americans” it makes their claim look restrained, realistic, reasonable. It’s still bullshit, but it’s a flavour of bullshit designed to appeal to investors who imagine themselves as more discerning, smarter, less gullible.
The best cons work on people who think they aren’t gullible.
“According to what Mercedes wants shareholders to believe” is, I think, what you meant to say.
Remember, every company working on making self-driving cars has a strong incentive to constantly claim that results are “just around the corner”. Because shareholders are fucking idiots, and they’ll put their money into whoever bullshits them the most confidently.
Here’s the basic problem with this solution as far as I can see: assuming we’re talking about a distributed public ledger blockchain, you haven’t described how the chain is secured.
The existence of systems like “proof of work” and “proof of stake” is based on the need to have some sort of proving mechanism for validator nodes. You have to solve the sibyl problem, or else someone can just run 10,000 copies of the validator software on one computer, submit enough votes for a false record that it overwhelms any competing votes, and thus create their own version of the chain - now authorized as the definitive and true version - where they get free energy for life because they’re so staggeringly wealthy in your new currency.
Distributed public ledgers only work if you insert a real world cost to validation. Basically, something of value must be committed or destroyed in order to authorize a validator node. Otherwise you have to authorize the nodes yourself, and now you’ve just reinserted a central authority.
So what is destroyed or committed to secure your chain? Assuming proof of work, it would be hardware and energy. People would be burning power solving increasingly complex and entirely meaningless math problems in order to be allowed to act as a validator. So now we run into the problem of incentive; why would they do this? In basically every public ledger blockchain that exists, the answer is that they get paid. Newly created tokens are given out to validators as a reward for their work. And, inherently, those tokens must be worth more than the cost of doing that garbage work in order for validators to actually benefit in any way. Without that, the incentives don’t work, and the validator nodes all shut down, destroying your blockchain.
This is why speculation and rampant deflation are inherent to cryptocurrencies; because in order for the validator system to not be overwhelmed by a single bad actor buying a tonne of computer hardware, the complexity of the validation (hence, the cost of the work in spent energy) must scale with the amount of hardware in the network, and that means that the cost of being a validator scales with the amount of hardware in the network. So as your network grows, the value of the token grows, or else the network dies.
But you’ve decided that people will also be rewarded with a token for the actual physical act of generating solar power and feeding it to the grid. And they’ll pay for power with those tokens. So your system is unbalanced. You pay people to generate power with newly created tokens, and then destroy those tokens when they’re spent to buy power. But you, presumably, also pay people to run validator nodes (because how else is your network secured?) using newly generated tokens, so you’re giving out more tokens than the actual amount of generated power in the system. That means you have too many tokens chasing a limited supply of goods.
So now you either have to allow people to overbid for power, creating rampant runaway inflation, or you have to keep the cost per kwh fixed, and create a situation where people go to get power but there’s none in the system, because you’ve got floating, “empty” tokens that don’t actually reflect a unit of power generated. And since you’re paying for power going into the system with these tokens, either way you’re destroying their perceived value and that means you’ve destroyed any incentive to sell power to your network in the first place. They’ll just sell to the grid instead.
And moving to Proof of Stake or Proof of Storage or any other proving mechanism doesn’t solve this problem, because ultimately they all rely on the validator committing something of value. If they don’t, it’s by definition no longer a proving mechanism, because the cost of sibyl attack becomes zero (or close enough to zero as to be meaningless). And if every validator must offer something of value, they must get something of value. Which means you have to generate tokens and give them out to the validators, and you have to ensure that those tokens have a real-world worth that is commensurate to the value that the validators commit.
Agreed, I would love this.