EpicAglet's recent activity
-
Comment on Meta is adding ridiculous ‘rate limits’ and a soft paywall to its smart glasses in ~tech
-
Comment on Reddit will require you to be logged in to use old.reddit.com in ~tech
EpicAglet Link ParentOr they aim to kill old.reddit.com but want to limit backlash, so they make it increasingly hard to access before removing it entirelyOr they aim to kill old.reddit.com but want to limit backlash, so they make it increasingly hard to access before removing it entirely
-
Comment on Reddit will require you to be logged in to use old.reddit.com in ~tech
EpicAglet Link ParentI only really used reddit these days if I land on it through google. The annoying things is that often, if you are looking for a discussion or an opinion etc, reddit and sometimes quora are still...I only really used reddit these days if I land on it through google. The annoying things is that often, if you are looking for a discussion or an opinion etc, reddit and sometimes quora are still is the only real options.
-
Comment on OpenAI says the US government will vet users of its latest AI model in ~tech
EpicAglet Link ParentWhile I share the sentiment, it got me thinking about this for a sec and reached the conclusion that I'd still rather have them not get too involved. The bursting of the AI bubble will not be fun....While I share the sentiment, it got me thinking about this for a sec and reached the conclusion that I'd still rather have them not get too involved.
The bursting of the AI bubble will not be fun. It'll probably be a massive economic crisis. I think it's true that the longer this goes on, the worse the collapse will likely be. And the great cheeto getting involved will most likely hasten this. So on the surface it seems like you'd want this outcome.
But maybe not. I'd fear he'd get himself into a situation where he'd feel like he cannot afford to let the bubble burst. I could see them somehow finding a way to use the US government to keep the bubble intact (for a while longer). What happens when it bursts might be much much worse then. The more involved the Tangerine Toddler gets, the more likely I'd think this would be.
Alternatively, if it pops too soon, the economic crisis will happen during this presidential term. Luckily of course, this master deal maker has shown that he is tremendous when it comes to dealing with a crisis, the best.
You might be better off with a slightly worse crisis if it is being handled by an administration that actually knows what they are doing. Over the bubble popping now, but then expecting Agent Orange to deal with the fallout effectively.
-
Comment on German retailer Thomann is suing Fender over recent cease and desists in ~music
EpicAglet Link ParentThanks. That makes sense. I suppose that's how they get you. As long as the inconvenience of switching remains greater than your annoyance towards them, you're not going anywhere as a customer. I...Thanks. That makes sense. I suppose that's how they get you. As long as the inconvenience of switching remains greater than your annoyance towards them, you're not going anywhere as a customer.
I picked Reaper simply because of the free trail and I liked the vibes the company was giving off. I kind of assumed all the major DAWs are comparable, but I never really looked into the others much. It seems like that is the case from your comment, so it's reassuring to know that I'm not missing out on anything by sticking to Reaper
-
Comment on German retailer Thomann is suing Fender over recent cease and desists in ~music
EpicAglet Link ParentA bit off-topic, but why do you use Studio One? I'm fairly new to recording, but I am currently using Reaper. I'm happy with it, but never really tried other DAWs apart from briefly playing around...A bit off-topic, but why do you use Studio One? I'm fairly new to recording, but I am currently using Reaper. I'm happy with it, but never really tried other DAWs apart from briefly playing around with Ableton in the past.
Does the DAW really make a difference, or is it just a matter of being used to a certain workload?
-
Comment on German retailer Thomann is suing Fender over recent cease and desists in ~music
EpicAglet Link ParentExactly. Yet they've been using this to send cease and desist orders to a lot of small little luthiers, in order to pressure them to stop selling S-style guitars and even destroy their current...Exactly. Yet they've been using this to send cease and desist orders to a lot of small little luthiers, in order to pressure them to stop selling S-style guitars and even destroy their current inventory.
It's seems however, that they made the mistake of also targeting some of the big players. Thomann, who owns the brand Harley Benton, is by far the biggest retailer outside North America. And they are the 3rd largest distributor of Fender guitars in the world after Guitar Center and Sweetwater. They have deep pockets and can afford to fight this out in court. For reference, Fender has a revenue of about 700M USD whereas Thomann is about 1.5B USD.
Moreover, they are based in Germany where the default judgement was won. So they will be very familiar with the legal system there.
Overall, this whole debacle seems like a giant mistake on Fender's side. They've lost a great deal of reputation in order to bully some smaller companies, and now they're in a fight with one of their most important distributors.
Meanwhile, a company like Fender relies largely on brand reputation in order to make its sales. These days, others make similar and arguably better guitars (I'm looking at you, PRS Silver Sky). People buy Fender because of the name on the headstock. What for heaven's sake were they thinking?
-
German retailer Thomann is suing Fender over recent cease and desists
25 votes -
Comment on Does generative AI have a natural limit without a major innovation? in ~comp
EpicAglet Link ParentI understood it as that progress is logarithmic as function of resource cost. So that's why progress seems limited despite massive investmentsI understood it as that progress is logarithmic as function of resource cost. So that's why progress seems limited despite massive investments
-
Comment on Drawbacks to Iceland having its own currency likely exceed the benefits, the country's Finance Ministry said, citing the conclusions of a government-commissioned report in ~finance
EpicAglet Link ParentIt loads again. So it appears that it was just down temporarilyIt loads again. So it appears that it was just down temporarily
-
Comment on Who’s buying SpaceX and Anthropic? in ~finance
EpicAglet Link ParentI think the massive valuation is because of a bubble, yes. One that will necessarily pop at some point. And I am rather worried about what will happen to the global economy when it does. I am not...I think the massive valuation is because of a bubble, yes. One that will necessarily pop at some point. And I am rather worried about what will happen to the global economy when it does.
I am not sure what you mean by buffer. I can imagine some companies might keep some money they raise by selling stock as a war chest for when their stock price tanks and they are no longer able to raise money, if that's what you mean.
My expectation is that the crash comes some time after the main IPOs. So you might make a fair amount of money buying their stock, assuming you sell before the crash.
The core issue is that AI is just not worth as much money as people speculating think it is.
When LLMs started to become commonplace, people thought it would be huge and be everywhere. That made it very valuable. People were right. That is exactly what happened.
But the promises were bigger. It would dramatically change most industries. A lot of jobs would be replaced by AI. And other jobs would see massive productivity gains. If so, companies would pay fortunes for access to AI.
Some productivity gain for certain jobs seems to be there. Some jobs can be (partially) done by AI, though often with a lot of human oversight. The promised gain just doesn't seem to materialise, which pulls into question how much these AI companies can really earn in revenue. The tech seems quite far from changing this equation fundamentally.
Meanwhile, we've reached a point where marginal improvements in the capability of these models requires astronomical infrastructure investments. And this scales terribly, in the sense that the next step up is much larger still.
It seems like the scaling cannot go much further anymore until the world simply cannot invest more. That's the problem with this kind of exponential growth. The money they raise by going public means they can stretch this limit, but it’s still there.
Basically, while they can probably increase revenue significantly by changing how they monetize but nowhere near enough to justify a trillion dollar valuation. Meanwhile expenses are unsustainable. And even if they stop the hyper scaling, expenses are very high. So they will likely not have margins that are too great.
So very large and valuable companies, but much less that what they will be worth at the IPO.
Before the AI craze, the big tech companies were trading at a multiple of maybe 8× or 9× revenue for some. But those are highly profitable companies with high margins and a lot of growth potential.
OpenAI for instance is now estimated at 50× revenue. And it seems to me that they probably will end up trading at a multiple lower than Alphabet or Meta once the dust settles, due to their bottom line not being great even if they become profitable.
-
Comment on Who’s buying SpaceX and Anthropic? in ~finance
EpicAglet Link ParentI have some faith in the long term success of Anthropic. But to me it seems that clearly anything AI is severely overvalued atm, including them. I'm pretty confident it'll be like the dotcom...I have some faith in the long term success of Anthropic. But to me it seems that clearly anything AI is severely overvalued atm, including them.
I'm pretty confident it'll be like the dotcom bubble. A lot of companies died, but also a lot of the big players are still around today. A handful eventually reached that valuation again, years after the crash.
I think Anthropic has the potential to survive if they handle it well. Yet they'll crash hard along with everyone else. And the entire global stock market will probably suffer too.
Anyone with exposure to the stock market will probably lose a significant amount of money. But the AI stock will crash the hardest. And others will suffer in the recession that follows.
SpaceX probably will not go away either, but I believe they are targeting a 2 trillion USD valuation. Their total revenue is 20 billion or so. That's clearly just the AI bubble inflating the stock price like crazy, based on some fairy tale of datacenters in space. They too will crash hard.
Plus knowing Elon Musk, he will do whatever he can to inflate the price and cash out at the peak. He does not shy away from manipulating the market
-
Comment on Drawbacks to Iceland having its own currency likely exceed the benefits, the country's Finance Ministry said, citing the conclusions of a government-commissioned report in ~finance
EpicAglet Link ParentThanks. I tried that though. It seems like archive does not want to load for me. archive.is and archive.today don't load either. I assumed it's down for everyone, but maybe it's only for certain...Thanks.
I tried that though. It seems like archive does not want to load for me.
archive.is and archive.today don't load either.
I assumed it's down for everyone, but maybe it's only for certain users. I believe that has been the case in the past
-
Comment on Drawbacks to Iceland having its own currency likely exceed the benefits, the country's Finance Ministry said, citing the conclusions of a government-commissioned report in ~finance
EpicAglet LinkI can't read the article, but I assume it is the finance minister arguing in favor of adopting the Euro on EU membership? Good. From what I understood, the currency and fishing rights were the...I can't read the article, but I assume it is the finance minister arguing in favor of adopting the Euro on EU membership?
Good. From what I understood, the currency and fishing rights were the main reasons Iceland put the negotiations on hold. So that would make it significantly easier to join
-
Comment on OpenAI is preparing to file for an IPO in the coming weeks in ~finance
EpicAglet Link ParentFor sure. Not only do the investors cash out big time, it is probably the only way to survive the crash. I can see two possible moves here. Either they cash as much as possible during the bubble...For sure. Not only do the investors cash out big time, it is probably the only way to survive the crash.
I can see two possible moves here. Either they cash as much as possible during the bubble to build a war chest for after the crash. Which you then use to buy up all the competition afterwards, so that when the dust settles you monopolise the market. Or they try to ramp up spending big time, bet on the bubble lasting for a while longer, and try to outpace the competition.
Or a combination of the two of course.
Though I am leaning to the bubble bursting sooner, rather than later. So perhaps option 1 is better then. Yet from what I gather, Sam Altman owns a lot of stock in companies that OpenAI does business with more than of OpenAI itself.
So I'd also not be surprised if he will go for option 2, to personally enrich himself even if it means running OpenAI into the ground.
-
Comment on A random sci-fi question for you in ~talk
EpicAglet Link ParentI was curious, so I did some digging. On the Martian surface, the eq radiation dose is 0.64 mSv (source). This would put you at about 1.6 Sv for those 7 years. For an adult, the cancer risk...Of course, at our present level of technology, so would a stint in the Mars gulag. Unless I'm misremembering, Mars doesn't have a notable magnetosphere, so anyone stuck there is gonna pick up a hellacious dose of radiation unless the colony is pretty deep under the surface. That raises further logistics questions.
I was curious, so I did some digging. On the Martian surface, the eq radiation dose is 0.64 mSv (source). This would put you at about 1.6 Sv for those 7 years.
For an adult, the cancer risk coefficient after exposure to 1 Sv is 4.1% (source (table 1)). So if we crudely linearly extrapolate we get to about 7%.
This is of course not accounting for any shielding you may have due to the habitat, but either way we see that cancer is still far from guarenteed.
Though in reality you'd also need to account the trip there. On Mars, the atmosphere and the planet itself block a big part of the radiation. So the travel there would be a significant chunk if the radiation dose.
Then again, in this hypothetical scenario we have the technology to travel to a far-flung earth-like planet. So maybe none if this is all too relevant.
-
Comment on Anyone else a bit unnerved by the number of visible satellites? in ~space
EpicAglet (edited )Link ParentEvery single car built after a certain year also has cameras slapped all over it and are connected to the internet. They might not be recording you, but they totally could do that. Bonus points if...Every single car built after a certain year also has cameras slapped all over it and are connected to the internet. They might not be recording you, but they totally could do that.
Bonus points if you realize that the Chinese government now owns a big part of some of these companies and have a lot of influence over them. So they could turn on an extensive surveillance network in foreign countries if they so pleased
-
Comment on The zero-days are numbered — Firefox team uses AI to find and fix vulnerabilities in ~tech
EpicAglet Link ParentPlus I can totally imagine that the kind of vulnerabilities that AI creates, are also the type that agents intended to find vulnerabilities will more likely miss. So overreliance on these tools...Plus I can totally imagine that the kind of vulnerabilities that AI creates, are also the type that agents intended to find vulnerabilities will more likely miss. So overreliance on these tools could create a situation where the security is full of holes that is easy for a skilled person to find. But who knows. We are still at a relatively early stage in that sense
-
Comment on Survey reveals almost 50% of California teachers may quit teaching soon in ~life
EpicAglet (edited )Link ParentI mean to some extent it is true that they just want to feel appreciated. And the best way to show employees they are appreciated is by paying them more.I mean to some extent it is true that they just want to feel appreciated. And the best way to show employees they are appreciated is by paying them more.
-
Comment on NASA chief classifies Starliner flight as “Type A” mishap, says agency made mistakes in ~space
EpicAglet Link ParentExactly, I meant stop giving them contracts for things like Starliner. Boycotting them entirely would likely be counterproductive and probably illegal due to how procurements work. But perhaps one...Exactly, I meant stop giving them contracts for things like Starliner.
Boycotting them entirely would likely be counterproductive and probably illegal due to how procurements work.
But perhaps one route would be to put more stringent QA requirements and frequent audits as part of the contract on anything mission critical or anything Boeing is expected to apply. With the threat of fines worth much more than the value of the contract in case they fail. But I don't know what the contract looks like now of course, and to some extend I'd expect this to already be there.
More drastic measures would be something along the lines of the government pressuring Boeing to sack all of the upper management. But for that they'd need sufficient leverage first. Right now that doesn't seem realistic.
They'd need to start failing to the point they need a bailout or something like that. Or attempt to temporarily nationalise the company by buying them out. Or have the threat of suspension from federal contracts, but from what I understand (not a lawyer) they'd need to show something like criminal misconduct to suspend them. Or perhaps somehow force them to split off the parts that do the government contracts. That might help contain the rot.
Meanwhile, both Boeing and Airbus still have long order backlogs for the commercial airliners. So Boeing won't really feel the consequences much there anytime soon either. Even if in the future Boeing jets become worse than Airbus' and airlines start to favor Airbus. They simply won't be able to scale production up overnight. And Boeing still has years of orders on the backlog that they'd need to burn through.
If Boeing does end up declining, it'll likely be a very slow process I reckon. Unless something changes the equations a fair bit.
So it's basically like BMW's heated seat subscription. A bit of a strange move considering the backlash BMW got, and right now they should probably be pushing mass adoption rather than monetization.