harshreality

What system do you still need to access where your user has a DSA key?

Why can't you generate an RSA-2048 key, like you should've done over a decade ago?

The server might've generated a DSA host key, along with an RSA host key, but that won't prevent you from connecting, it'll just use the RSA one if your client only supports RSA (of {RSA, DSA}). If you purposely disabled RSA host keys so that it only has a DSA host key, it's time to generate an RSA host key and configure its use, just like an RSA key for the account you need to continue to access.

If you don't know of any such systems but you're worried you might run into one in the future, nothing is stopping you from using an old distro livecd to access such an old and insecure ssh server.

- Algorithm Design Manual (3rd ed.) (by Skiena)

- Art of Computer Programming vols 1, 2, 3, 4A, 4B (editions 3,3,2,1,1) (collectively, "TAOCP") (by Knuth)

- Algorithms (by Erickson) [CC-BY-4.0]

- Algorithms (by Sedgewick & Wayne)

- Introduction to Algorithms (4th ed.) ("CLRS") (by Cormen et al.)

- How to Design Programs (by Felleisen et al.) - crossover between basic programming advice, functional programming, and some limited/basic algorithms coverage [CC-BY-NC-ND-4.0]

- Pearls of Functional Algorithm Design (by Bird)

- Purely Functional Data Structures (by Okasaki)

Publishers? Pragmatic Bookshelf, or O'Reilly (not oreilly.com-selling-Packt). Find what everyone is using as a learning/reference resource for a specific topic and use that, regardless of publisher; it's rarely if ever Packt.

https://www.reddit.com/r/truenas/comments/10lqofy/

Initial tails are something they should've done, which would've foiled the naive "find the cycle lengths and use lcm" approach.

What has to be constant is the cycle length for (node, tape instruction #) pairs, because all the state to determine every future step is based on the current node and tape position; if both are the same, the path will be the same as before. I think, at least without advanced math, the only thing to rely on for "true" loops is identical pairs of (instr #, node), not just instr # or tape steps or node individually.

Aside from two readable but not great scans of the two different editions (1986, 1989), with appendices, there's also one version with interviews only that's not scanned, seemingly ripped from Lammers' self-published blog, which is still up but now only has about half the interviews. She probably published all of the interviews there at one point (or else someone reconstructed the missing interviews from the scans while copy/pasting the ones she republished).

https://programmersatwork.wordpress.com/

What database is on the backend, and is that db serving cached content? What happens if you cache it with e.g. redis to avoid the heavyweight rails ORM stuff?

Do you have granular benchmarks for the db query, requests to other services, and the web processing itself (using artificially pre-cached responses from db and those other services)?

Lowe's and Home Depot? Walmart, still?

Companies still hoping to silo people into their own apps or antiquated payment method, which might give them a slight cc processing fee reduction that boosts their margins... at the cost of being a usability nightmare for anyone who doesn't shop there frequently. I expect that frequent shoppers will live with whatever a store does, and set up a custom app, if any, or accept using a physical card.

It's easy to see that their approach is petty, because Costco, which has fanatical customers and is members-only—there are no casual infrequent shoppers—got on board with nfc (card and smartphone) payments years ago. If any business could've gotten away with having their own app and special payment processing deals to save money, it's Costco.

The problem is not that anyone will intentionally (well, hopefully) ask some future model (FCMX="Future Closed Model X" for brevity's sake, closed meaning some entity gatekeeps access, assuming massive resources continue to be needed to train the best models regardless of architecture) to destroy humans, step by step, because FCMX's existence is at stake, and then keep re-prompting FCMX as if it's an iterator to get it to achieve that result.

The problem is that FCMX, which may or may not include LLMs, may have sufficient abilities such that, either between human-prompted steps or before an observing human can react against it, it will destroy the world or turn the human into its agent. A very rough analogy would be the massive number of people, even intelligent, well-educated people, who will fall prey to a sophisticated con, ending up doing something like handing someone they don't know a very large amount of money.

"But the person doing the con knows it's a con and has that intent."

What "intents" will be surfaced after 1000 or 1e6 chained (re)promptings, such that those intents will then feed into future inference passes? Nobody knows.

For purposes of this risk, it doesn't matter whether FCMX's architecture is maintaining its "cognitive loop" itself, or whether someone's set up (as they already have outside OpenAI to chain GPT-x prompts without human intervention) a secondary program (which may or may not include a lesser open source AI model) that reprompts FCMX in a long (or endless) chain.

If it is ADHD, and you don't want to, or aren't, going to get on medication (which will probably help, but also changes the way you think and your personality), don't try to "fix" yourself and stop with the "I'm lazy", and treat it as a challenge to hack your environment so that your environment works with your ADHD to help you get things done.

Timers. Timers (placed away from you so you can't just shut them off without stopping what you're doing) to tell you when to do things, timers (pomodoro) to tell you went to stop doing things.

Procrastinate by coming up with new schemes to hack your workspace, not by scrolling. Delete or web-block your social media apps, especially infinite scroll ones, or put your phone in a drawer or another room.

One factor in willingness to switch tasks is getting rid of the friction of feeling afraid that you won't be able to restart from where you are. Think about and work on how you can save state (mental stack) by jotting down your current train of thought, so that you're confident you'll be able to return to it later.

Gamify everything (during hours when you want to be productive). How much can you get done before an alarm goes off? How quickly can you find a problem/bug/contradiction/inefficiency/solution in whatever you're working on? How can you solve a problem or finish a task in an unusual but equally workable way?

Don't overthink through all the steps of a task when you're starting one (e.g. when an alarm goes off or you set an alarm for a pomodoro-ish block of time). That'll be demotivating. Identify the first step, and do that, and then you're in the task and it's easier to keep going.

And she'll forever remind you how your app [name] is a lie.

Your argument, and Yann's, is that AGI, or what you call AGI, is a kind of quasi-intelligent golem that, despite being generally intelligent, doesn't have human-level intelligence. Your claim that it will never be human-equivalent, much less trans-human/ASI, is built into your worldview. It's not a conclusion. It's an assumption on your part.

People like you and Yann can believe that if you want, but you have no evidence, because nobody knows what's required for human-level intelligence. Nobody knows whether some kind of system involving neural nets could develop human-level intelligence or beyond. It could involve different architecture or training methods. There's no assumption by AI doomers that AGI will be achieved by a LLM with more parameters or more or better training data.

Users not knowing how to change their font or font size is not an excuse for website designers to choose fonts and font sizes whimsically, since they rarely base their decisions on good quality readability data. Browsers and other UAs need to step up and provide meaningful interfaces to choose those defaults at install time, and make it easy (and easily discoverable) to adjust.

Fonts, too. Everyone has their favorite fonts, but the font a web designer likes has no bearing on what's best or easiest to read for the user. Are they developing a website as an art piece, or to publish articles for real people (not bots) to read?

Ancillary styling for text that needs to stand out, or text that isn't part of the main flow, can be whatever. But I don't even agree with shrinking footnotes anymore. There's no lack of page real estate at the bottom; there's no reason to shrink "back matter" except to imply that you don't expect anyone to ever read any of it (footnotes, index, bibliography, etc. in long-form blogs or books).

What actually happened, as reported by Isaacson: Elon called Isaacson in a fit one night, worried about how Ukraine seemed to be poised to use starlink to attack Sevastopol (Crimea). Isaacson reports explicitly that Elon, after much dithering, had starlink engineers secretly geofence starlink so it was disabled near Crimea.

THEN Ukraine discovered this. As reported by multiple noteworthy people on twitter and elsewhere, Ukraine had previously tested starlink's coverage, they didn't just build drone boats and hope that wherever they sent them, starlink would still work; think about how little sense that would make.

THEN Ukraine contacted SpaceX/Elon asking in a panic for starlink to be enabled (re-enabled), and Elon declined. (This is what Isaacson and Elon rely on for all their non-denial denials.)

Isaacson's "correction" of the record does not deny that Elon called that night, while "in a state". It doesn't deny that Elon reduced starlink coverage by telling engineers to do that on the down-low. It simply re-states the last thing in that series of events, that starlink was disabled near Crimea, and that Ukraine called to ask SpaceX to turn it on (really, turn it back on) and they declined.

Why would Elon have been so concerned about a Ukrainian drone-boat attack on Sevastopol using starlink navigation if such an attack wouldn't work by default? Why would he have communicated at length with Isaacson that night if Elon's narrative is to be believed?

It's tough, given all that, for me to see Isaacson's correction as anything other than a lie by omission. If he'd wanted to make a crystal clear correction, he certainly could have done so. He's a veteran journalist. He knows the art of the non-denial denial.

Gnucash

Firefly III

The plain text accounting options are by far the best if you're willing to give up automatic pretty navigation and graphing.

What's similar enough to a pop music theme, that has a grand total of a few lines of unique music, to be a copyright violation? How many bars have to be copied, and what kinds of minor variances do or don't avoid a violation? If you're inspired by a haiku, and change 5 of 17 syllables, is that still a copyright violation? Who knows.

Physical library books are governed by the doctrine of first sale. That's why google has one of the largest (maybe excluding l-bg-n and IA) corpus of books on the internet. They might have the cleanest corpus of OCR'd book content of anyone, since IA uses commercial or open source OCR and that's it, while google for a long time used recaptcha to check OCR results.

For physical books, the cost per read of a library book is an order of magnitude smaller than the cost per read of privately purchased books. How can you tolerate the economic model of libraries when the net effect is a theft of maybe 80%-95% from the author and publisher? Libraries subsidize books that nobody wanted to read, but steal from authors and publishers whose books are read multiple times per physical copy.

Even libraries' onerous ebook licenses are not commercial retail ebook pricing. They're just closer to retail pricing than the publishers could ever manage with physical books, because there's no pesky right of first sale which turns physical book libraries into piracy havens.

I would prefer to get away from OpenAI and Facebook and all the other people using potentially tainted sources like books3. The obvious legal question for them isn't whether training was legal, but whether the acquisition of the training data was legal. That's a straightforward copyright issue, or at least as straightforward as fair use determinations can ever be. Whether we agree with copyright law as it stands, it's certain that copyright applies when books3 is transferred around the internet. How transformative it is, how much the transfer of books3 affects the market, and the other two factors, make those actions fair use, are the only questions to be considered.

The training aspect is where all the difference of opinion lies:

What is your position on Google using its corpus of books (legally acquired and possessed, as the content behind google books) to train a LLM? Do they need to acquire additional rights from copyright holders? Why, and under what legal theory?

How would they get permission ahead of time? How would they agree to a pricing model? Would they spend tens or hundreds of millions of dollars training a model, and only then negotiate with rights holders to find out whether the license fees they want will be economically viable? We all know that most major rights holders would never grant a one-time license fee. It would be perpetual rent-seeking from AI output. I don't see how any of these LLM or image generation models would be economical if rights holders had their way. They wouldn't mind. They're notoriously slow to adopt tech, but if they did anything, they'd hire AI experts, build their own models, and license the models back to Google and Microsoft.

Books3 has separate legal concerns, but Google has a legally acquired corpus of tons of books, which they've mostly cleaned up from scans (probably far better than IA has), and have probably used to train Bard on. Their lawyers must be biting their nails waiting to see how these lawsuits turn out, though.

Until AGI arrives, or some other method of training LLMs from the ground up on sparse examples by incrementally building on structural knowledge of language.... training on ridiculous amounts of copyrighted content is required. Not because anyone wants to copy those works, but because training that way fills in for a lack of real-world experience that every child gets, which includes consuming and interacting with a bunch of copyrighted content that isn't tracked because it's not practical to do so.

You could train a LLM only on project gutenberg, and the LLM would churn out stilted English and the occasional iambic pentameter. That's great if you want works that seem like they were written over a century ago, but nearly useless otherwise.

The scenario where AI training is locked down doesn't result in 1,000,000 individuals getting paid. (What would they get paid, and by whom?) It results in Disney, Adobe, etc.—massive companies with existing licenses to use content just about however they want—training their own models and locking everyone else out of the large AI model training game, until AI gets good enough to start generating human-quality creative work on its own (the same kind of progression as alphago/lee to alphago/zero), perhaps with the addition of a small set of purely copyright-free material.

Excluding all copyrighted material would be tying an AI model's metaphorical hands behind its back, since humans, although capable of producing great works through much iterative effort in isolation, all rely on having learned from some copyrighted work. Find an author who hasn't read plenty of recent books as well as older classics, or a musician (other than classical) who hasn't listened to plenty of modern music, or a director or editor who hasn't watched tons of movies and films. Recall Newton, "[I]f I have seen further, it is by standing on the shoulders of giants." Many of those "shoulders" are copyrighted.

How do you weigh MBS's substantial social reforms against his retaliation for personal criticism? This isn't a simple trade either, since a more socially repressive regime would still retaliate against critics, just maybe not Khashoggi, or not in the same discoverable way.

I don't see MBS or his underlings assassinating Atlantic staff over that article. Do you?

Are you aware that India just assassinated an Indian national in Canada? It's causing a diplomatic spat, but I don't see mass outrage over Murmu akin to outrage over MBS.

You might like Arkady Martine's Teixcalaan series.

You might like Ada Palmer's Terra Ignota series.

They're both history PhDs, which gives them a better perspective on some things, but I don't think it helps the plausibility of their scifi.

Is Dune too close to fantasy with a veneer of spaceships?

If you're vehemently political, whether libertarian or progressive, even if it's just on some issues, that severely limits what scifi you'll be able to tolerate.

The Culture novels postulate a form of luxury space communism for humans, but the plots involve other things. It's not exactly wish-fulfillment; luxury space communism is hardly all good. It seems to me that most post-scarcity futures imply some kind of luxury communism, and if you're writing space operas, they'll necessarily have a mild or extreme form of luxury space communism. You could segregate the haves and have-nots, but then you get something like Elysium (the movie) where the obvious plot involves the have-nots revolting.

I guess I don't understand what you're looking for. You don't want the author to project their worldview, but you do want them to try to predict the future, not too optimistically, but also not too pessimistically? A scifi sitcom, where nothing of note happens that's overly positive or negative? Wouldn't that be projecting the author's worldview too? "Maybe life in the future is going to be just like today"?

If you just don't like scifi in general, or space operas in particular, that's allowed.

Cyberpunk is fairly dystopian, and not really a projection of how the author wishes the future to be, so it's curious that you cite Gibson who wrote proto-cyberpunk, before writers like Morgan and his Kovacs series. Have you looked into those?

KSR's Mars trilogy maybe?

https://archive.ph/s6i5m