Yes, you could have done similar with GIFs already with Netscape 3 and IE 3 (I think, IE 3.1, or was this just the Mac version?). "Dynamic HTML" with z-layering became a thing with Netscape 4.0 and Internet Explorer 4.0, which became available in Summer/Fall 1997. With this, relative positioning and proportional text (at variable character width) wouldn't have been a problem.
(With the NS/IE 3 generation, you would have been limited to lines of prepositioned fixed-width characters in a table. So, not really applicable for this.)
I think, the decisive factor may have been badly managed CRTs tending to display greys with a somewhat reddish hue (esp. on Windows, not so much on Macs with a screen gamma of 1.8) and JPEGs generally enjoying better color support.
(JPEG was displayed at sRGB or at any embedded IEC profile, which wasn't available for GIF, where the effective color was pretty much hit and miss. So, if you had HTML text with a defined hex-color, a GIF of the same color wouldn't match, while a sRGB JPEG would. Matching JPEG and GIF colors would have posed a similar problem. A reddish hue on grey looked kind of cheap and brownish text was probably the last thing Apple would have wanted.)
Yes, this is pretty much how we did such things already in the 1990s. The problem with this is that this only supports a predefined, rather limitted character set. But you could have expanded on this and may have loaded custom characters on demand (which would have been a rather small payload for a single character.) But there's yet another problem, namely color management, which isn't supported by GIF (which we need for the transparent background), and they probably really wanted to go with JPEG for this. (At that time, CRTs were still much around and those tended to show greys with a reddish tint when managed badly. Something Apple would have certainly wanted to avoid.)
BTW, here's a PoC (a website from 1998 written in about two hours [so not much of hassle even then] demoing pure JS rendering from a data file, all just classic table layout rendered in the browser, all styled text and the ticker are rendered from single-character images): https://www.masswerk.at/demospace/relayWeb_en/welcome.htm
The first paragraph kind of reads like an inverse of Kant's criteria for a moral action: are you sure that there was no enjoyment, no pleasure was involved in this action or resulted from it? At the same time, it kind of establishes the ethics of self-entertainment, and by this, maybe, an imperative of joyless joy? It kind of echos 19th century 'accomplishment', a modest self-entertainment by the fruits of virtuos learning and training, while with kind of a consumerist twist, as this now involves the use and "modest" development of technology, which introduces an external, social dependency, which it at the same time tries to eclipse.
Meaning, I think, besides utilitarian one-time, one-purpose scripts, programming is essentially an at least potentially social activity, there's always a potential social horizon to it. Just like with any other activity that aims at a work.
PS: For the those seeking comfort in literature, we get some discussion on accomplishment in the work of Jane Austen, as in a worthy and virtuous entertainment for young gentlewomen, accomplishment for display, "true" accomplishment, as an ability to entertain oneself under conditions of intellectual austerity, means of intellectual independence, etc. Not to the least in "Pride and Prejudice", where accomplishment is discussed as a sibling and foil to Darcy's male conception of inner moral virtue, which, in fact, amounts to an socio-economic Regency scandal, as he entirely ignores his first-born duties in good-natured impulsiveness, bad decision making and short-sighted expenses, which an entire middle class depends on. (Is there a similar side to accomplishment?)
Notably, this was written at a similar time, when fortunes made from trade and a new kind of availability of manufactured goods was turning the (paper mache) table, in terms of what may be a good and fulfilled life and a worthwhile activity.
> "It's not clear what exactly the end goals are, as the information appears to be publicly accessible anyway."
Another reminder that LLM productions are really a prompt on us to inflate this output with meaning. (And that LRHF is really the engineering that makes this likely to happen.)
(I.e., all the webpages with a rotating globe with "News" spelled around it in the fashion of the animated Universal logo, embedded as an animated GIF [GIF89a].)
I loved this one as a kid in first grade because it just makes it so clear and obvious what the problem even is! I didn't need diagrams pretending to inflate a balloon in a cylinder, or an animate gif warping from globe to flat.
Yeah -- I remember in old (1960s) TV shows and movies the headquarters of organizations would always have a map like that, often huge on one wall. It makes sense -- why not just peel the Earth like an orange and flatten it out?
It's the great desemantification machine: want to look, read, sound, feel just like the median, without personal traits or individual expression? This is for you! (And people used to talk about communism and everything being same. Well, tech-broism actually achieved this.)
I think, there's a systematic flaw in RLHF: constructs reinforced as effective or sophisticated style will always suffer overuse and significant overexposure. Which will be also very obvious when used out of the intended context, it's copied from. There will be always a "smell", something that causes us to turn our backs to these texts, and possible the supposed authors, as well.
There's also the problem that certain types of phrasing are perceived as effective, because they mark a pivotal point in the progress of the text and are, as such, used sparingly, but are now becoming everyday templates that incorporate whatever is available in the context. There is no way this passes the Turing test of a competent reader.
And there's yet another issue: in social research, there has been the concept of semantic position, indicated by deviation from the mean (or median). If you don't deviate from the mean, your semantic position is zero. There's simply no expression. In this sense, next token prediction really amounts to a desemantificiation of the context. There's really no sense in uttering any of these productions, no plausible motivation, other than for the purpose of raising you hand to be seen.
> The Ecclasiast looked under the sun, but there was something he didn't understand. Something that wasn't right. Something that was not as it was supposed to be. And here is what the Ecclesiast didn't understand. Here is what nobody understood. Not then. Not in the years that followed. Not now. It was not the swift who won the race. Not the strong who won the battle. Not the wise who earned the bread. Not the men of understanding who gained the riches. Not the men of skill who gained the favor. And here is what I found: to any story of success, there is an element of unpredictability and chance.
(There are really just two possible outcomes: either the article is right or in, say, two years, we will be all writing and talking like this, as in humans learning from mediamatically reinforced human feedback.)
I already notice LLM speech patterns in people that use them a whole lot. If you speak more than one language I recommend talking to LLMs in a language that you don't use when speaking to people.
I have a tendency to switch between different languages when interacting with LLM depending on the language of the most likely sources the LLM will pull. It's interesting because the llm speech patterns vary from language to language. And I am much faster at noticing LLM generated text in English (which I use 70% of the time) than in French or Spanish.
In Greek they are still atrociously bad, still feels like they translate an English thought. Nevertheless, the etymological insights they can give on words or Ancient Greek passages is making reading the ancients really fun.
The most frustrating one is when it does the papering over analysis and adding pushback to seem helpful. No, there shouldn't be a spot "where you pushback" unless it makes sense and responds to someone's concern!
Like with an LLM I would rewind the conversation. What do I do, shame them?
I am a bit of a luddite in this domain and have so far managed to resist the lure of using the generator to expand my thoughts, and I still catch myself writing "it's not just an X it's a Y" and other generator type tells. If it infecting my patterns it is totally entering the wider subconscious as "How to write" (Sighs)
That's the flaw in RLHF: constructs tagged as effective or sophisticated speech become templates, are increasingly used out of context (e.g., "not just A, not just B, but C" is usually meant to provide some synthesis and further the progress of the text) and suffer from significant overexposure. (And gone is the em-dash…)
It's bad in and of itself the way LLMs use it. They separate it into a weird declaration when it's usually something you'd say at the beginning and/or end of a paragraph or passage where you're making a case for that statement.
It's part of how LLMs often seem to repeat the same thing over and over again. Most arguments are X is Y (and the others are X is not Y, or X is likely Y), and LLMs lean on "X is Y" because it is the most distilled form of that argument - a thesis minus the meat. Also "This is not a Y, this is a Z."
They're just thesis sentences inserted in inappropriate locations. The reason it makes them overdramatic is because separated from the argument, they seem like they should be premises. Conclusions without premises have the feeling of demands.
The question is: how did your test reader respond to the version without the "It's not just an X, it's a Y." ? Perhaps the discussion went:
Test Reader: I got confused when you started talking about X as though it were a Y.
Author: But it is a Y. I should flag that up, where I introduce it.
Alternatively, the test reader comments
The villains monologue wasn't unhinged enough. He says that making himself universal dictator is democratic because he gives people what they secretly desire. I feel he would express himself with more empty rhetoric.
Author: Agreed. I'll add "It's not just democratic, it's not just super-democratic, it's ultra-democratic!!!"
Test Reader: Scary!
The error lies in using "It's not just an X it's a Y." for emphasis rather than to ward off a likely misreading.
That one starts with a glaring hallucination "he didn't understand, nobody understood, not then or now...". But the second part is IMO better than the Ecclasiast, which has outdated phrasing ("nor yet riches to men of understanding").
> When I look under the sun, it's not the swift who win races, nor the strong who win battles, nor the wise who earn bread, nor the smart who gain riches, nor the skilled who gain favor. Ultimately, it's the lucky; every contest has a degree of unpredictability and chance.
The construction of meaning is totally distorted, though. In the original, the first part is entirely motivated by the conclusion, and puts emphasis to this by a crescendo of counterintuitive observations. This is what made this memorable phrasing. In the version suggested here, there's just an incompetence attributed to the original author, underlined by repetitive phrasing, so that the supposed speaker may claim all the glory by uttering an observation, which is in and by itself rather banal. You wont think about this twice. Thus, the entire construct ends in a whimper (so it may fit the template) and there isn't really any reason to put it that way, or to say this, in the first place. Its perceived purpose may be rather to occupy space and time (at the cost of the recipient.)
Of course. As we're told to prepare for the new times… :-)
On a more serious note, I think, if we're going to criticise the kind of distortions LLMs do to language and the construction of meaning, we should be able to come up with similar productions, just to demonstrate where and how this goes wrong.
(Much like the early Wittgenstein pointed out some constructs and declared, this is not a meaningful sentence, in fact, it's not a sentence, at all. And went on to define the limits for what a meaningful proposition may convey and what it can not, what we may expect and hope for in an expression and what we may not. In other words, is there even enough in the toolset of LLMs to construct meaning? Is traversing a semantic field in an apparently well-formed sentence already meaning or is it not?)
> two years, we will be all writing and talking like this
I have seen kids talking like youtubers: the top 10 things I like, or what I need to do before I die (the kids is 9 years old), ... not great (like repeating ad tunes or slogans) but not as bad as AI bullshit-talk.
My only hope is that AI bullshit-talk is so boring and predictable that it has low impact in our speech. And looking at what other commenters say boring and insufferable seems to be the case.
(With the NS/IE 3 generation, you would have been limited to lines of prepositioned fixed-width characters in a table. So, not really applicable for this.)
I think, the decisive factor may have been badly managed CRTs tending to display greys with a somewhat reddish hue (esp. on Windows, not so much on Macs with a screen gamma of 1.8) and JPEGs generally enjoying better color support. (JPEG was displayed at sRGB or at any embedded IEC profile, which wasn't available for GIF, where the effective color was pretty much hit and miss. So, if you had HTML text with a defined hex-color, a GIF of the same color wouldn't match, while a sRGB JPEG would. Matching JPEG and GIF colors would have posed a similar problem. A reddish hue on grey looked kind of cheap and brownish text was probably the last thing Apple would have wanted.)
reply