fafferlicious

There's definitely hope that it synergises with what's going on due to the steam deck.

Personally, I can't wait to buy two controllers and the machine. The flawless experience of the deck is amazing. And because it's Linux, I'll just install YouTube, jellyfin, any app as a non steam game and I'll have the perfect smart tv appliance.

Stream games, play games, run any program I want through steam big picture - I can't wait to bury my Shield.

I'll never have to connect a tv to WiFi again. I'll never see a fucking ad for anything on my TVs home screen again. With KDE connect my phone is a remote. I'm so fucking pumped.

This is bullshit, IMO.

It's profitable. It's just not as profitable as being a douchebag with business decisions - like enshittification and planned obsolescence. The moment you try to to publicly fund your company, you suddenly have a fiduciary duty to your investors to maximize returns. And let's face it, appliance manufacturing has high up front costs that will likely require an IPO to satisfy your initial private investors.

Just because you make quality appliances doesn't mean you can't make money maintaining and manufacturing parts for quality appliances.

Power is a function of the size of the difference you're testing for. Idk how to really calculate it accurately, but this site would suggest to discriminate between 7.2% and 7.4% you'd need over 300,000 participants

I'm kind of skeptical of that, but I don't know enough about clinical trials to know that.

I'm curious if you have recommendations on how to structure or keep the notes. I find that I struggle reading technical documentation or how to structure notes so they're easy to refer to. Have any tips or guides you can share?

Do not mistake the opportunity cost for inviability. Billionaires are not lining up to build one with their money may be a fact (I don't know), but even if it is, it is entirely possible that the rate of return is just too low for them to consider investing in. Why make 3% EV YoY when you could get 5% EV YoU? Almost double the rate of return!

Not so easily proven.

It's really not some mythical user, man. I just laid out the two main methods we were comparing of updating firmware. One process requires literally 0 user intervention after device setup. The other require the user to intervene. In what world is the process where the end user has to do literally nothing to get updates inferior to any process that requires intervention? It's not that the device has Internet that's the issue. It's that there's nothing respecting privacy! We're on the same side but you wanna...what...make it illegal to do OTA updates because internet connectivity for devices is intrinsically invasive somehow? I just don't get it. Your position doesn't make sense.

Make an app

And now the app pulls gelocation data of the user and takes the data on the device and uploads it through the app because no privacy laws.

Why does using a USB stick to flash the firmware equate with chewing your own food, but using an app to update the firmware not? They didn't compile the app from code!

there is no compelling use case for constant internet connection for these types of appliances

I literally just gave you one. The ability to push out OTA updates for bugs greatly reduces the complexity of maintaining technical support and development to support legacy features. I don't know, maybe instantly patching a critical Bluetooth vulnerability to protect users privacy via OTA updates is compelling enough ?Do they have to do it that way? No. The device could just get it via BT from the phone app. Basically the same thing as the device itself doing it.

But the privacy issue still remains even if we remove WiFi from the fucking vacuum and then just let the phone app have WiFi access instead.

Did I make my point better? Removing internet access to the device, doesn't remove the privacy concern. It just moves it somewhere else. And yes I know there absolutely is a way to do firmware updates in some privacy respecting way, but if you maintain that a 0 step process is less preferable to the majority of users than any process that requires multiple steps... Then I don't know. Interact with users more? There's plenty of boomers peck typing their way around the internet.

I admit you have points, friend. I share your concern about privacy. But come on... We got where we are in part because the general users don't give a flying fuck or even think about privacy.

The sheer fact that internet access exists for them indicates there's a use case. And it's obvious. Over the air updates are objectively a better user experience. Maybe even the best. For the average consumer that doesn't even know to use ctrl + c, what do you think they'll choose: dragging out the USB stick/cord or just having the device take care of itself?

The problem isn't that the devices have internet access. It's that we don't have appropriate legislation governing protection of privacy that would keep this kind of snooping in check.

I'll give this one go just on the off chance you're being genuine - since you seem to have some sense of rationality to your position.

You're wrongly underweighting the damage that will get done while the courts correct things. Yes, eventually the courts would restore broadcasting licenses. But there's appeal after appeal after appeal. That could kill the stations even if ultimately they'd win. And don't say that the licenses couldn't be revoked in the first place. They did it to fucking science funding. Funding mandated by Congress just stopped, and Congress did nothing. Sure it's coming back now. Kind of.

We know how this plays out already. And we can't even count on the courts. The supreme Court has shown a total disregard for their duty with some of their rulings.

Like the executive order on birthright citizenship. Before the supreme Court the government didn't even argue the legality. They said nationwide blocking orders were the real problem. The Court agreed.

Or the use of race or language as a cause for ice detention. Blatantly against the forum amendment and the Supreme Court said it's fine. They're snatching US citizens for fucks sake.

This is all easily searchable. If you really care about being rational and reasoned, you'll put in the effort. This isn't people desperately trying to be wrong. These are people that have seen what is likely to happen. They've updated their priors more accurately than you.

People don't sympathize with their trash when it ends up in landfill.

We all have our hard lines, and I don't fault yours.

However for the scrollers passing by, they have a lifetime membership of about ~$45 USD. I found it worth it. It was super easy to get used to, and I honestly can't stand any other launcher now.

I like it a lot and would hate to see it die out, so it was easy to justify for me.

I didn't make a strawman.

The criticism of "just probability" falls flat as soon as you recognize current expert consensus is that humans minds are... predictive processors....

They're just like us!

Except....

Where LLMs struggle adapting to things outside of distribution (not in the training data) they do not have a way to actively update their weights and biases as they contextualize the growing novel context.

You wanted to attack LLMs underlying principle of being probabilistic word sequence generators. But that's it. That's what they do. They have no understanding outside the context of word order to know that typically if a sentence starts "The quick brown..." the word fox frequently follows that phrase. Therefore, a fox is probably quick and brown. And if something is quick and brown, it might be a fox. LLMs are not intelligent not because they rely on probability.

LLMs are not intelligent because they do not know anything. They repeat patterns in observed data. They do this in an intentionally leaky way to generate new sentences it hasn't seen before based on context it has seen them in before. Any reference of "thinking" or "learning" is just anthropomorphism or an inaccurate and misleading (though useful) approximation. They have no concept of "correct." It's why you can bully them into agreeing with you. They're dumb.

Look, I'm not going to get any more into this because you used a lot of big, jargony words without any context. Words like "normalize to the tribal opinions", "RLHF", "intermodal dissonance", or the biggest offender "confabulations." Those would only be used by a person more knowledgeable in the field or a self-fashioned intellectual trying to flex.

If you're an expert, I offer advice I got in grad school: speak to your target audience. Unfortunately, I can't engage with most of what you said because I frankly have no fucking clue what you're saying.

You all are missing the forest for the trees! LLMs are just like us in how we think! We're all just probability generators! No, they can't think or reason beyond known data sets. Yes, they fail at extrapolating information which is the basic component of reasoning. But you guys don't get it! They're just like us and smart!