d110af5ccf

> mostly in the dwindling twilight of their cognitive ability

Have you actually read SCOTUS trial transcripts? Because they come across as incredibly sharp people to me even when I don't agree with their views.

> A lot of modern psychology suggests they can't.

Actually there are psychology studies examining professionals who have to deal with bias (scientists, judges, etc) that indicate that the competent ones tend to slightly over rather than under compensate.

> Law needs a larger tool chest that can make argument based not on precedent, but on what exists right now in American society.

How is that different than legislating from the bench? Isn't updating laws a job best reserved for the legislature?

> Same people that complain in this post about over-jealous verification, will complain in another post about misinformation and propaganda.

A bit tangential but actually I suspect those are nearly disjoint sets. In my experience the people who complain about misinformation and propaganda are okay with identity verification and censorship while those that want privacy (such as myself) typically dislike censorship and don't want a central authority getting involved to judge whether something is misinformation.

It largely comes down to trust in authority and centralized versus decentralized system design.

Nope. All physical effects are bounded by the speed of light. (As far as anyone knows, anyway.)

The only weird one is quantum entanglement, but even then information transfer doesn't travel faster than light and that's about all I know on that subject.

> but displaying success messages even if it's not

Please don't do stuff like this. Some of use use VPNs and aggressively block third party JS. These sorts of shenanigans render many sites unusable which is frustrating but simply not using those services is a price I'm more than willing to pay at this point.

This provides some detailed low level information. https://wiki.osdev.org/System_Calls

> Wouldn't it be better to design a protocol that doesn't have those concerns in the first place?

I'm referring to design concerns, not usage concerns. To minimize privacy concerns related to usage, you first have to add privacy as an additional design concern.

And privacy is a particularly difficult one because it means a lot of different things to different people. Can you reuse connections at all? What about content fingerprinting based on transfer rates - do you need to add cover traffic? Voice codecs, for example, will inadvertently reveal when the stream goes silent because they more or less stop transferring data. Instead of designing voice codecs to be less efficient, it probably makes more sense to optimize them to the maximum extent possible and then handle privacy concerns at a higher layer on an as-needed basis.

Separating design concerns to the extent reasonably possible tends to yield significantly less complex end results which has a whole host of benefits. A Swiss army knife often proves to be sub par in practice when compared to more specialized tools. Do one thing and do it well.

> A good design would be easy to implement without opening up users to security and privacy issues.

Simple designs are nearly always easier to implement. Compare the code base of OpenVPN to WireGuard for example. I think it's quite reasonable for a spec to accommodate implementations that want to make various optimizations that sacrifice user privacy. In cases where that's a concern implementations can omit such optimizations as necessary based on their specific usecase.

> no more paying comprehensive insurance

Assuming you can afford to replace it on your own if you need to, this one really adds up.

Doesn't it make more sense to separate concerns? Build the highest performance protocol possible that can operate efficiently under a wide array of conditions and then route via an overlay such as i2p whose sole job is to provide privacy if you want to do that.

Also isn't an implementation free to avoid things like connection reuse if it so chooses? For that matter some browsers support fully containerizing things per site visited.

> "AI" can't "escape" its programming

I don't think you can back that claim up. ML models as often used can't escape their programming. But I don't see any fundamental reason you can't loop one back on itself.

Isn't this already done to some extent by using a model to assess its own output versus the current environmental state? Recurrent models combined with reinforcement learning actually look a lot like what you're claiming can't be done.

> There is no mental model of the world to rationalize against

There's a theory and some associated research about a model that performs well necessitating an internal model that accurately reflects the environment in which it operates. Or something like that. Sorry I can't recall the paper off the top of my head.

> Whatever tasks a ML model has been deliberately programmed with

Yeah, no, that's not the case. See the entire field of reinforcement learning.

If you're going for shelf life I've eaten canned chili that was more than 15 years old and it was fine. Most things seem to change texture in unpleasant ways sitting in a can that long but chili doesn't. Not that it's viable to live off of in terms of calories and storage space, just an interesting observation about shelf life.

Starch and cellulose are both glucose polymers. The only difference is the way in which the glucose units are attached to one another. Existing enzymes can only cleave one or the other. Ruminants don't actually have the necessary enzyme, but instead rely on bacteria in their stomach to do this for them.

anyone can cash a fraudulent check, though it may be against their "rules"

edit: Quite puzzled that this is being downvoted. The comment I responded to is suggesting to disregard the TLD's rules when registering domains. You certainly can do that but as far as I understand it is a form of fraud (you are materially misrepresenting yourself and your situation to the TLD). Realistically probably the worst that will happen to you is that your domain will get yanked if they figure it out but that might well be a problem for you and regardless it doesn't change that what you did likely violated multiple laws.

> Is there an easy and objective way to tell the difference?

Better question, does it matter? As you said, bad hires are expensive. We aren't talking about language barriers here, we're talking about behavior that comes across as (your words) capital A asshole. If the person is going to behave that way on the job and it drags the team down, does it matter why they behave that way? I know I wouldn't want to be stuck working with them.

> training in private flight schools and then working as flight instructors themselves

Who attends those private flight schools? Unaffordable aircraft obviously won't improve things and seem quite likely to make things worse.

> we had net neutrality for years and it never improved the reach of broadband into rural areas

The issue there isn't net neutrality but rather open access. Countries with affordable internet generally have open access requirements for network operators.

Net neutrality is a separate issue to do with ISPs attempting to double dip, leveraging a large customer base in order to do so.

> Netflix pays for connectivity just like consumers do.

Yes, Netfix pays their ISP. They should never need to pay the customer's ISP. Peering should be cost neutral. ISPs shouldn't be allowed to double dip.

From GP: "extreme amounts of resources"

"Extreme" here meaning ... the connection bandwidth that was explicitly paid for? Except for the fine print where the ISP notified the customer that they were going to horribly oversubscribe things. It amounts to fraud as far as I'm concerned.

> Can they selectively drop packets

Absolutely not. That practice should literally be illegal.

> or do they have to drop them randomly?

Yes. This is the only reasonable approach.

> And can you pay to be in a “never drop my packets” group?

No. That should be illegal to offer.

If the ISP's network is incapable of meeting their customer's demands during peak hours and they oversubscribed themselves so badly that they're contemplating such tactics then they ran their business dishonestly. They fraudulently made promises that they were incapable of delivering on and got caught. They need to fix their network or go out of business. And possibly should face legal consequences for their behavior.

> Prioritizing non-streaming-service (e.g. throttling Netflix at prime time) packets could, for example, help a regional ISP stay competitive.

No. There are lots of things companies could theoretically do to be more competitive. That doesn't mean those things are good or should be legal.

> That's different than the anti-competitive behavior the big ISPs engage in when ...

Actually I think the two are tightly coupled. There's a Hetzner article linked to elsewhere in this comment section that does a good job of articulating it. https://web.archive.org/web/20170607120440/https://wiki.hetz...

> With a cost-neutral peering, each network operator pays for the expansion of its own network itself. Likewise, every network operator shall bear the costs of its own router interfaces at the connections between the networks (peering points).

> Large content providers have a strong interest to reach the DSL customers of DPCs with the best possible performance, and therefore are looking for alternatives to congested interfaces between the carriers. These are found usually through direct connections of the major content suppliers to the respective DPCs. At this point, the DPC can use their market position, since as a result of congested interfaces no other carrier can reach the DSL customers of the respective DPCs with sufficient speed. Therefore, the DPC can set prices, free of competition, which interested content providers have to pay to achieve performant connctions to "its" DSL customers.

So effectively ISPs are incentivized to neglect peerings that they aren't currently double dipping on in order to use their customers to coerce payment. Note that (unfortunately) DTAG was eventually successful in coercing Hetzner to pay for peering with them. They did this by abusing their customers. It's textbook monopolistic behavior.

And? When I last checked sbcl didn't support it. Are you implying that has changed?

Neither did sbcl support cross compiling. To the best of my knowledge all you can do with sbcl is save an image.

When I was looking in to this I read through many discussions where the developers appeared hostile to the idea of implementing these features. A common refrain I encountered was that tree shaking is unnecessary because the core images are compressed.

I thought the kernel had the ability to restrict DMA on modern devices. But a sibling comment indicates that modems integrate GPS support these days so I guess it's irrelevant.

> To do this work in software would require a workstation.

These days hardware capable of that throughput can be had for less than $500 and draws less than 50W. 3.9 gigabits really isn't all that much by modern standards.

You don't. As others noted, you can save a core image. Cross compiling, tree shaking, and other such things don't exist in the sbcl world and (from what I've read in the past) the people developing it have little or no interest in those features.

If you need those sorts of features paired with common lisp you could give embeddable common lisp a try. Otherwise if any lisp language will work for you then chicken scheme and gambit scheme both compile to c.

Many sites including HN (https://news.ycombinator.com/rss) provide RSS feeds, so you could presumably navigate to the relevant page for the share/comment/like functionality. I realize that's not ideal but it seems well within reason to me.

Software exists to translate RSS feeds to ActivityPub actors (ex https://github.com/dariusk/rss-to-activitypub). Going the other direction, Pleroma provides RSS feeds. Not sure about Mastodon or Misskey.

> it will strain Wikipedia's governance model and become more of a single point of failure

Why can't the Wikipedia model be adapted to a federated, directly community run approach? This works well enough for services such as email, matrix, and the fediverse. There's gravitation towards centralized hosting services but that's largely behavioral - the model itself works perfectly fine with lost of small players.

Heavyweight multimedia can be a challenge but text content itself is quite easy to serve up from very small devices.

No, GPT-3 still produces gibberish at times. The majority of the good examples still ramble like a schizophrenic person. Much of the output is uncanny, interesting, and impressive in its own right but I wouldn't describe it as human level.

DALL-E 2 is different from what I've seen. The things it produces seem to actually make sense the majority of the time. The outputs are strikingly similar to what a competent human might output as opposed to one with a severe mental illness.

I'm sure part of this is an inherent advantage that DALL-E enjoys regarding context. Art is supposed to be artistic whereas text is expected to maintain long distance logical consistency of abstract concepts across a stream of output and also to communicate something concrete. So in a sense the bar for art is probably lower in many ways.

> We need this if we want to keep communicating and exchanging ideas with strangers online.

I disagree. There are indeed a few cases where you can't initially be certain that you're interacting with a human but sustained interactions over a long enough period of time solve that. So the only time we would "need" such a approach is for people to know that something written by a total stranger was written by a human. But you should be getting to know people and not trusting complete strangers anyway!

Even then it doesn't actually solve anything. Paid actors can still post things in an attempt to influence complete strangers. This is exactly what happens right now. So you'll know the paid actor is a human. Or at least paid off a human to provide access to one of their online identities. Great.

So it's a knee jerk response to scary new ml but the ml isn't actually that scary once you think about it and the proposed response doesn't actually solve the supposed issue or any other current problems and it erodes our freedoms in the process.

You seem to be describing hosting your own service, changing the underlying hosting provider, and tying your identity to ICANN DNS.

There are two separate things here, identity portability and data portability. If your identity is portable (say a key pair) but the service hosting the data goes down and you don't have a backup or there's no way to import the data to a new provider then you've got a problem. Similarly, if the service goes down and you do have a backup of your data and a way to import it to a new provider but your identity isn't portable (perhaps it was user@domain and you don't own the domain) you also have a problem in that case but it's a different problem.

And neither data nor identity is the same as account, at least the way I was using the words.

> Who would host the data?

Anyone who wanted to. If you wanted to make sure your personal data remained available when you turned your computer off or that it could be downloaded quickly then you'd need to arrange for paid hosting. Pick up cheap vps and share it with your family and friends. Whatever.

> how would that hosting mechanism be resistant to collusion and manipulation

It doesn't need to be, that's the point of using strong cryptography.

> it seems you will need some sort of decentralized state/data tracking (like a distributed blockchain ledger)

No, you just need to keep track of keys. Your identity is your key. We already do this. Certificate authorities, ssh keys, etc. The issue you're describing is one of using a centralized authority to solve the problem which is hardly the only solution and certainly doesn't require a blockchain.

> my backup account is only going to allow me to keep participating, but it is not going to save the conversations I had on the lost one.

The same issue exists for other services including email, irc, and matrix. The obvious simple and immediate solution is to retain copies of things you care about locally instead of expecting a particular service to stick around forever. With email this used to be standard practice before the rise of webmail. Even now most desktop clients sync to and retain a copy of everything on the local device.

> > Accounts aren't portable but there's no inherent reason that accounts need to map 1:1 to identities.

> Again, you are talking about identity portability, which is not the same as "creating one different identity for each different server".

It seems like it is though? Servers need a way to handle authorization so they presumably need accounts. With a simple protocol, accounts map 1:1 to identities as far as the software and protocol are concerned. If you make a new account then you also make a new "identity" from the perspective of the software. The obvious way to solve that is to map any number of accounts to one (or possibly multiple) identities, where an account is the thing the server is concerned with and an identity is something else such as a private key.

There's an additional question about the mechanisms for account creation and authentication. Maybe you don't want to maintain an email and password based login for a long list of accounts that are each tied to the same backing identity. There are multiple approaches available there as well, from oauth providers to public key based authentication possibly including things like cross signing.