It really isn’t a different debate when you’re talking about putting them on the blockchain, and all that other engineering has already been done by other distributed social networks.

Trust, consensus, and access control are session-layer issues that don’t need to be solved by a transport-layer protocol. Social networks deserve to be able to forget things.

We already have that, it’s called a Distributed Hash Table, no blockchain required.

No worries, I’m merely confident that the tradeoffs necessary to employ a blockchain aren’t worth the supposed benefits thereof.

What if we don’t want global usernames? What if we’re entirely satisfied with global user IDs in a DHT?

Seems inefficient, couldn’t the same thing be accomplished using local DBs rather than the world’s most inefficient ledger?

It was a Saturday, but I was on-call when Networking shit the bed. One of the main trunk lines degraded and they took almost five minutes to switchover to backup 'cuz their automated degredation monitoring was on a five-minute interval. XD

Oh heck, that sounds promising. 😺

The TOR network is indeed the most widely-used implementation of onion routing, but it isn’t the only example.

My go-to reference is Retroshare, an open-source app that implements onion routing on top of an encrypted friend-to-friend network:

http://retroshare.cc/

You only connect to your trusted friends, but by passing messages along the Kevin Bacon chain it’s still possible to reach practically anyone on the network. Retroshare’s built-in services include email, instant messaging, traditional web forums, microblogging, and Reddit-style karma-ranked forums/linkboards, and third party plugins include voice and video chat. It’s desktop-only, but I think it demonstrates that serverless social networks are possible.

Depends on the usage.

“Social Networks” can be a reference to the various social media services available online, but they can also be an reference to the collection of social connections one has offline.

Ah, that’s my bad. Didn’t bother learning who the CEO was. XD

Fixed my post, thanks!

Say you want to send a letter to a friend of a friend, but you don’t know their address so you can’t send it direct.

Instead, you can package your letter inside a second letter to your mutual friend, asking them to finish filling out the mailing label for you so your message will reach the intended recipient.

They call it “Onion Routing” because the message can be wrapped in multiple layers of these routing requests, with each recipient stripping off the layer addressed to them and forwarding the remainder on to the next connection in the chain.

Using this protocol, so-called “Friend to Friend” networks can still enable communication between non-friends so long as a “5 degrees of Kevin Bacon”-style connection exists between you and whoever you’re trying to reach.

She’s right, but centralized services like BlueSky won’t be it.

Social networks don’t need servers playing middlemen, friend-to-friend networks with onion routing to pass messages to friends-of-friends would more accurately reflect the structure of real-world social networks.

The difference being that googling the problem and visiting a page on stackoverflow costs 50-500 times less energy than using ChatGPT.

If your problem can be solved by a bot, then an old fashioned touch-tone phone menu would be an entirely sufficient solution, no “AI” needed.

If not, then plugging an LLM into your IVR will never be worth the expense since the customer will need to talk to a human anyway.

“AI” is a bubble. Sure, it might have some niche applications where its viable, but it’s heavily overpromised and due for disinvestment this year.

Companies are going to waste money on it, drive up costs, and make the consumer pay for it, causing even more unnecessary inflation.

The “AI” bubble will burst this year, I’d put money on it if I had any.

The last time we saw a bubble like this was “Web3” and we all know how that turned out.

“AI” is a bubble. A lot of these concerns will go away this year once the bean-counters do the math and realize that the benefits of running generative neural networks aren’t worth the costs.

A single chatGPT query costs about 50-500 times as much energy as a pre-Bard Google search, to say nothing of the engineering time needed to build the models. And, since LLM outputs can’t be trusted, the end users will still need writers and developers to go over everything and check for hallucinations.

The trajectory here closely mimics “Web3”, when people thought that massively redundant distributed ledgers were going to be the next big thing, despite the fact that traditional electronic ledgers beat the blockchain in literally every aspect of performance, efficiency, and security.

Soon, “AI” will be just as synonymous with “plagirism” as “cryptocurrency” is with “scam”.

Generative neural networks are the latest tech bubble, and they’ll only be decreasing in quality from this point on as the human-generated text used to train them becomes more difficult to access.

One cannot trust the output of an LLM, so any programming task of note is still going to require a developer for proofreading and bugfixing. And if you have to pay a developer anyway, why bother paying for chatgpt?

It’s the same logic as Tesla’s “self-driving” cars, if you need a human in the loop then it isn’t really automation, just sparkling cruise control that isn’t worth the price tag.

I’m really looking forward to the bubble popping this year.

It already is in the more expensive cities like Denver or San Francisco.

I don’t trust any company that still relies on easily-tapped fax lines whenevery doctor or tax office I’ve used has had a secure online portal for uploading documents for years.