ineptech

It's a speech by the author Robert Anton Wilson discussing the work of Alfred Korzybski, who helped to explain and popularize the map/territory distinction and how it relates to the philosophy of science.

I'm a huge fan of Wilson and Korzybski, the ideas Wilson discusses in this lecture definitely changed my life and the way that I think about science, but even I would caution that it's not all that relevant to the discussion here.

> When Congress certifies the election they do not publish that certification onto any blockchain. Maybe some day they will.

The Oracle problem isn't the fact that this doesn't happen today, it's that it can never happen in a way that is trustable. When Congress (or anyone else) does decide to publish the election results to a blockchain, every dollar bet on the outcome will be a prize to be won by anyone who can subvert the publication process.

> The only idea I see is for some certificates handed out by government to citizens and I absolutely hate it even in a democracy.

I think this is the answer. Governments already have the infrastructure to verify identities in person, and no other organization is going to build it.

The problem of moving information about the real world (stock prices, weather report, did the Knicks win last night) into a crypto ledger so that it can be speculated upon (e.g. a smart contract that only pays out if the Knicks win).

I don't get how anyone can call something with an eternally-fixed quantity a currency. What's the plan for when it starts appreciating and people stop spending it and start hoarding it?

I don't buy this. I'm sure most iphone users don't care when you ask them about privacy or manifest v3 as an abstract concept, but remember what happened when Apple tried to push a U2 album to them? They lost their collective shit. They may not write blog posts about privacy or donate to the EFF, but they have deeply personal relationships with "their" phone and they absolutely hate being reminded that it isn't really theirs.

If this weren't true, Apple could just start inserting ads into every iphone's Safari window tomorrow, and Youtube could serve the ad in the same stream as the video to defeat adblockers, and they'd make a bunch of extra money with no downside. The fact that they don't do this suggests that Apple and Google understand this: people only tolerate restricted platforms that do a convincing job of pretending to be unrestricted. In practice, this means that step 1 of Google foisting off user-hostile stuff on us is getting Firefox to include it too, which is presumably why they spend so much money on it.

I don't know about cool, but if my username dot com is amusing enough to qualify I've been curious to check it out.

More than ten years of consistent, substantive updates. Never had an expansion or DLC or "season pass" or in-game purchases. Best gaming $15 I ever spent.

I watched (okay, skimmed) the demo (okay, ad) and from what I can tell, it's a phone with no screen that you clip to your shirt and interact with verbally. The demo includes receiving a phone call, language translation, summarizing a meeting, and looking up a product (by UPC?) to see if the speaker is allergic to it. I assume it's OpenAI on the backend due to the speaker mentioning how excited Sam Altman is about it.

I can see what they're going for, if I had an always-on device hoovering up my data I can imagine an AI using it to do cool stuff. But the only way I'm wearing a 24/7 body cam is if the data is going to open source software on hardware I control, and AFAIK this is neither, so no thanks.

Not exactly what you're describing, but this is along those lines: https://slatestarcodex.com/2015/06/02/and-i-show-you-how-dee... It's pretty silly but the ending is both satisfying and relevant.

Sure, maybe? We (or at least I) know little about the actual problem here, and metaphors only go so far. But to my mind, "too many things trying to handle a request" gets a cool name because it is a fairly narrow and unusual problem, whereas "too many requests" goes by many names (DoS, hammering, flood, etc) because it's depressingly common.

It's the difference between one customer asking a hundred cooks for a waffle and a hundred customers asking one cook for a waffle. The former is the thundering herd (a bunch of processes trying to do something that only needs to be done once, causing resource contention) and this is akin to the latter (with the "customers" being parallel requests from the frontend).

There is no authoritative body on the definitions of words. More generally, if genocide can mean "not killling, but very bad" then it is not much use except as an epithet - a negatively-loaded bomb to be lobbed in partisan debates at people who you think are doing something very bad. Virtually every controversial policy could be described as, "Causing serious bodily or mental harm to" some group.

I understand where you're coming from, but -cide implies killing.

> financial speculation ... provides more accurate pricing and lowers transaction costs for the actual users of the futures contracts who take delivery of the commodity.

I understand and agree, as I've said pretty explicitly in both of the comments you responded to. I'm arguing that the finance sector should ideally be smaller than it is, but you're responding as if I said it should disappear entirely.

> In my opinion, your arguments are coming from an emotional place. Try and examine futures markets from a place where you aren’t thinking about greedy rich Wall Street guys, the amount of money they make is irrelevant to futures markets being useful tools for producers and consumers of commodities.

I don't have any an animus against speculators; please try to read more charitably. The rapid growth of the finance industry over the last two generations is a result of the policies we've enacted, and the position that it should be smaller is an argument for different policies. Similarly, "it seems like we are pouring [too much] of our resources and brainpower in to designing exotic new ways to bet on the corn harvest" is a complaint about the system that incentivizes that outcome and the policies that produced it, not about the individuals acting within that system.

> I’m not sure what you’re trying to imply...

I wasn't suggesting that commodities are similar to housing in any way; I was saying that, since a lot of people seem to recognize the societal cost of speculation on housing specifically, that I was surprised to be downvoted for complaining about the cost of speculation more generally (which, I remind you again, is not the same as saying it shouldn't exist, only that our economy ought ideally to produce less of it).

Sure, I get this and agree, but price discovery and facilitating markets are subject to diminishing returns just like anything else, right? I don't think I would've been downvoted for saying something like, "It's a problem that it's more lucrative to speculate on existing housing than to build new housing, so we should make regulatory changes to address that" and this feels analogous to me.

Was this not the whole point of the study in TFA comparing false memories of crimes to false memories of non-crimes? The assumption that people would be less willing to admit to a serious crime than to having been injured?

> Intriguingly, the criminal false events seemed to be just as believable as the emotional ones. Students tended to provide the same number of details, and reported similar levels of confidence, vividness, and sensory detail for the two types of event.

I'm not sure how this is related to my post so perhaps I was unclear. I'm not talking about individual corn farmers and the choices they make, I'm talking about how we as a society and an economy allocate our resources. I'm saying that derivative financial instruments have value, for the reasons I suggested and the others described by sibling commenters, but that the finance sector is larger than that value warrants.

I'm not sure why I'm being downvoted, as I didn't think this is all that controversial. Historically, finance was a much more boring and less lucrative field than it is now, and consequently much smaller. "I'm a super smart 18 year old and I want to get rich, so obviously I should go into banking" is a relatively recent phenomenon. I agree with everyone else here that the industry has value, so presumably its recent explosion in size has brought some additional value, but it's very hard to believe that value is large enough to offset the opportunity cost of a generation of ambitious geniuses not going in to science or industry or becoming entrepreneurs.

This isn't false but it feels reductive. A financial instrument that allows one to bet on the corn harvest is obviously valuable to the corn farmer, as it allows them to use profits from good seasons to hedge against bad seasons. They're also valuable to people whose business is affected by the corn harvest - cereal manufacturers, say. The problem is that they can also be used by people with no exposure at all who simply want to bet on the corn harvest, and from the scale of the finance sector it seems like we are pouring a lot more of our resources and brainpower in to designing exotic new ways to bet on the corn harvest than we are on growing corn.

> We need to rethink this model.

This problem is inherent to client/server software, and there are really only three ways to do it:

1. The server side of client/server is centralized and run by corporations

2. The server side is decentralized, meaning everyone has their own server

3. Abandon the server, clients connect directly to each other without a server intermediating

Option 3 would be ideal, but would require significant technological advances - it'll be a lo0ong time before bandwidth is cheap enough that Kim Kardashian can serve photos and movies to all of her fans direct from her phone. Option 1 is what we have now, and is terrible in a variety of ways.

Option 2 would be hard but is not obviously impossible, so still our best bet - sure, it's not viable now, but it sure seems like it could be, if an iphone's worth of r&d were put in to it. I would honestly be amazed if no one at Amazon is working on such a thing, since no one would benefit more than AWS from a future in which a cloud VM becomes one of the things that most middle-class families rent monthly.

I'm not a policy analyst, but this doesn't seem very difficult. Market failures caused by information asymmetry can often be improved by mandatory disclosure of some kind. Even if you disagree, the inability of a bunch of programmers to describe an effective regulatory solution doesn't mean one doesn't exist.

I think this misses the forest for the trees; the handful of cases where they broke the law are nowhere near as deleterious as the rest where they didn't. I can't think of a better time for new regulation than a hedge fund extracting home equity from desperate people.

Yeah, "simple" as in we didn't have to make something that can learn general concepts, and then teach it language. It feels like a hack, doesn't it? Like you were working on a problem you thought was NP-hard and then you stumble over a way to do it in O(n^2).

> You cannot make accurate predictions without understanding the world that gave rise to these words.

I think you must either admit that chatgpt does exactly this, or else give up our traditional connotation of "understand". Chatgpt has never seen a sunset, felt pain, fallen in love, etc, yet it can discuss them coherently; to call that understanding the world is to implicitly say that the world can be understood solely through reading about it and keeping track of which words follow which. It's amazing that generating text from statistical relationships about tokens in a corpus, which generated nonsensical-but-grammatical sentence fragments at smaller scales, can expand to concepts and essays with more compute, but it is just a difference in scale.

> I’d be careful calling it simple.

I'm not calling it simple, I'm calling us simple! I'm saying that chatgpt is proof that natural language processing is much easier than I previously thought.

I have been thinking along the same lines - the fact that something as comparatively simple and barebones as an LLM can manipulate language symbols well enough to carry on a conversation suggests that it's a lot easier than was previously imagined. I used to think that language was one of the defining characteristics of intelligence, like the instruction set of a cpu, but chatgpt seems like persuasive evidence against that.

Answered upthread: https://news.ycombinator.com/item?id=35615264

Sure, maybe? I don't have a clear idea of where consciousness begins. If I were trying to make a conscious LLM, I would use a sort of feedback loop of automated prompts talking about its input in realtime. Like, if you showed it a picture of a tree, it might auto-prompt itself with, "What do I think of that tree?" and after it answers itself, "Why do I think that?", etc, and use the responses to update its own weights. I can imagine something akin to a personality developing out of that, and those automatic prompts sure sound similar to the "inner voice" in my head.

But what I meant was, even though I don't know exactly what's necessary for consciousness to exist, in humans or computers or dolphins, I feel like LLMs are just obviously too simple to be it. I've had a casual interest in NLP since the 90s and have followed AI and NLP progress to some extent, and up to a few years ago, I had this vague notion that language processing might be so complex that you need a human-like consciousness to wield it. I think chatgpt is the ultimate refutation of that. It's proof that, whatever part of our brains processes natural language and converts our stored knowledge of the world into words, it can run standalone - it doesn't need eyes or ears or thoughts or a childhood or some digital version of hormones or a thalamus or anything like that. Just input data and an algorithm to store relationships between tokens is enough.

That's a really interesting result, philosophically. Think for a moment about how we retain the information we take in. Consider, e.g. a college student who reads a dozen books about WW1, and then writes an essay about it a month later. It's not realistic that they can remember every sentence and every statistic; to retain that information, they have to condense it, summarize it, notice trends, and presumably forget large parts of what they read and don't retain. That is more or less what we call learning, right? Well, what chatgpt tells us that either that that learning process is entirely unnecessary to answer questions about WW1 based on books about WW1, or that the way a gpt turns training data into weights is roughly analogous to it. Either way, that's a surprising result. And (to bring this back full circle) I think it's surprising for the same reason many people thought Searl's Chinese Room thought experiment was unrealistic: the assumption that, since the process by which we turn input into stored knowledge and thence into a series of parseable tokens in English is hideously complex and mysterious, that a computer program that can do the same will need to be complex and mysterious as well, so complex that we can't rule out the possibility of consciousness. Whereas in reality, you can look at chatgpt and say, "No I pretty much understand exactly how this works, and you can tell it's not conscious because it's not using any cpu."

I may be talking to myself in a dead thread, and am certainly rambling, but it helps me sort things out and develop my understanding - very much like that loop of self-prompting I mentioned earlier :)

It's only a mistake if you take it too literally and forget it's a model, e.g. "I measured the electron to have a mass, but I can't see how an excitation in a field would have mass so I guess I must've measured wrong."

I'm not sure what you're talking about, but it's not very related to my point about "is", which was that describing things in "[noun] is [adjective]" terms carries the implicit assumption that the universe is made up of things have objective qualities, which is a useful model most all the time but still a model, and hence ultimately wrong.

Fake edit: this is why it's helpful when discussing quantum mechanics to banish "is" from your vocabulary. For example, the debate over whether an electron "is" a particle or "is" a wave is much more tractable when you switch to debating whether it behaves like a particle or a wave in some situation. In college, I found this very similar to how it is useful when learning relativity to avoid the word "simultaneously".