For anyone as confused as I was: yes, this is indeed a link post on lemmy.world pointing to an article on kbin.social hosted by kbin.projectsegfau.lt and ultimately linking to social.bbc.

The old Fedi switcharoo.

"Good vibes only" seems to be embedded in the culture of web development today. Influential devs' Twitter accounts have strong Instagram vibes: constantly promoting and congratulating each other, never sharing substantive criticisms. Hustle hustle.

People with deep, valid criticisms of popular frameworks like React seem to be ostracized as cranks.

It's all very vapid and depressing.

Thanks for this. I skimmed the proposal doc itself and didn't quite understand the concern people have with it – most of the concerns that came to my own mind are already listed as non-goals. The first few lines of this comment express a realistic danger that's innate to what's actually being proposed.

Just tried the demo yesterday. The tutorial's integrated into the gameplay in a way that didn't feel obstructive to me. It's less like an old-school sandbox tutorial and more that the game makes it obvious what you have to do for the first mission. And it seems to focus on the new mechanics since the basic stuff is already made obvious by overlays showing the controls.

There will be people who have no capacity for nuance and see this as a boolean thing, and for them: the tutorial's not skippable, no. But for most people, it shouldn't be an issue.

This got me to look up iFixit's guide to Switch battery replacement. It's better than some of my devices, but as soon as a replacement involves spludgers and adhesive it crosses a "yuck" line for me, going from something that looks kinda fun to sort of dreading that I'll break it.

For contrast, past Nintendo handhelds made this a doddle, even in the post-AA era: here's the New 3DS battery replacement guide. The DS Lite even had a little battery door.

I only had occasional luck with this even on Reddit. Some smaller subs for hobby stuff had genuinely good advice, but a lot of times it'd just be people repeating the same brands and products with a shallow recommendation. And there was a lot of astroturfing. Over the years I've learned to ask elsewhere:

For major appliances, the best approach I've found is to find a local business – a well reputed one that's been around for years, and does service as well as sales – and simply ask the salespeople what they recommend. If the shop's willing to warranty it, it's probably good enough.

For gadgets I tend to start looking at recognizable review sites that are easy to skim (RTINGS is especially useful, but Ars, The Verge etc. all have decent reviews) and then expand out to YouTube for the products I'm most interested in. Sometimes it's a good idea to look up the company itself for anything that might change your mind about them (Western Digital's unlabeled change to SMR drives is a recent example).

Shoes and clothes are the hardest thing to get good advice on. The most useful advice I've received has been very general stuff about what to look for in fit and quality. I've also found that high ethical standards from a clothing company tends to go hand in hand with quality and longevity.

Cars are an area where Reddit was still helpful. YouTube can be helpful here, but not so much typical car review channels: the most helpful YouTube videos are often from people who've owned a particular model for a year or so and can speak with experience about its quirks.

Finally, and most of all, I've learned to check the instinct to look up reviews. It's worth spending some time to research stuff between you and the ground, or that you'll use daily, but I've wasted too many hours comparing details that really don't matter. Make sure it's something you legitimately care about before you reach for other people's opinions.

The Verge article is paywalled for me, but the screencaps Alex shared in his toot don't really support his summary. The article mentions that Threads can import content from Mastodon as an example of the sorts of things ActivityPub supports, and that's about as close as it gets.

And then there's this:

The company is planning to create a roundtable for administrators of other servers and developers to share best practices and work through problems that will inevitably arise, like Meta's server traffic putting strain on other, smaller servers.

Emphasis mine. How would Meta's server put strain on other, smaller servers if it's not federating with them?

I'm fully willing to believe Meta wants to EEE ActivityPub, but this particular claim doesn't seem to check out.

Maybe this is just a contrarian view, but I see "AI" as a potential rather than a technology. Right now, transformer-based technologies are what most of us mean when we talk about AI, and it's not clear to me how much more potential that idea really has. When I look at how much energy it takes to set up something like GPT-4 I see us pushing hardware to its limit and yet the outcomes are still too often unsatisfying. Significant breakthroughs are needed somewhere in that architecture just to do the kind of things we're trying to do today at the fidelity we expect and without breaking the bank.

The technology we have today might be to AI what the phonograph was to audio recording. As a technology we hit the limits of its potential pretty quickly and then… we fixated. Entirely different technologies eventually led to the lossless spatial audio experiences we can enjoy today, and seem more likely to carry future potential for audio too.

In that analogy, GPT might just be like someone arranging 8 gramophones in a circle to mimic the kind of spatial audio experience available in some headphones now. Impressive in many ways, but directionally not the path where potential lies.

Blockchain technology hasn't contributed anything of lasting value, and too much money, energy, and good will has been burned by people trying.

Its most popular applications are cryptocurrencies, which are used for gambling, money laundering, and for collecting payments from ransomware victims. Someone once bought a pizza with them, but since that time their transactions have become too slow and their value too volatile to exchange them for anything so concrete.

Various attempts have been made to use blockchain technology for public or shared databases, but it turns out to be worse than all the other faster and much simpler existing solutions in that space.

Others have attempted to bolt it on to various business and social systems, but it hasn't provided any practical benefit there either. It remains a slow and cumbersome alternative to every problem.

Its unique superpower is that it can be used to make contracts between parties that have no trust in one another and no social or legal system of enforcement, so long as your definition of a contract is sufficiently narrow, can be reduced to terms understood by the world's slowest logic engine, and is perfectly encoded the first time around and doesn't require any adjustment thereafter. If one or more of those conditions fail, you'll find yourself turning to the social and legal systems of enforcement you thought you didn't have.

This reply's interesting:

How can data licensed under the CC-BY-SA licenses (that SO content is licensed under) be "misused"? The license explictly allows others to do essentially anything they want with the data as long as attribution is given, in particular profit off of it.

When SO content is applied as parametric knowledge I'd expect the outcome to fail both the "BY" and the "SA" clauses, since model interpreters can't provide attribution for it and their output won't share the license. That's true even if output is considered public domain: CC-BY-SA content can't be moved into a public domain equivalent license. It seems practically indistinguishable from using any other in-copyright content as training material.

None of that's to say SO is right to stop data dumps. It feels like they're trying to find a technical solution to a legal problem, perhaps even one that rises to criminality on the part of Open AI and others?

I agree with your parenthetical strongly enough to rule out a typo.

This announcement lists things that Reddit will humor, for now, and as a way of cheaply outsourcing niche and difficult problems. It clarifies that everyday third party apps were never intended to have a future with the platform. They're simply an obstacle for Reddit's most convincing path to revenue.

I might even have forgiven Reddit if it had said so up front, but the story they've been trying to spin – with prices that just happen to be orders of magnitude in excess of anything devs might afford – is outrageously insulting. I've never had my trust in a brand demolished so thoroughly so quickly.

It's a fascinating idea, but it's reading a lot into a small effect on a very small study group.

The authors also don't seem to dwell much on the effect that novelty itself has on study outcomes. This is a common problem with A/B testing: introducing any divergence from what's familiar can draw participants' attention and affect outcomes. "Like" and "Dislike" buttons are familiar for anyone who's spent time on the internet, but "Trust" and "Distrust" buttons are novel and will be treated more carefully. If they were used widely then we'd eventually develop a kind of banner blindness to the language, and their effect on discernment would be further weakened.

This approach could also overindex comments that express risk and uncertainty. Well-worded concerns and calls for "further study" are a time-honored way of disrupting progress (never forget the Simple Sabotage Field Manual) but often sound trustworthy.

Which makes this comment rather ironic.