pilgrim0

The shape of the solution only becomes obvious once you treat the system as a product and define the requirements honestly.

The “honest” part is the important one. It’s also the hardest. It’s so tempting to project preferences onto a new development, which tends to confuse the distinction between desire and necessity. There’s so many things we want to try. Maybe having lots of personal projects to vent the need for novelty helps one to make good choices when it’s critical that they must be right.

The path of virtue requires we recognize and appreciate the grace that lies in being humbled by our own stupidity.

It’s because of cowardice. Discussing emotions entails being exposed to some degree, but this strain of so called rationality seems rather to be about distancing oneself from the subject in question, which tends to make people feel and behave like arbiters. One could argue that the poets are the real rationalists, at least the good ones. Because there’s no distancing from the truth, just raw and pure content, which is radically different than second order, cheap, hopelessly biased analysis.

Blender does open in a blink, it is scary how good it is. I share the exact same perception as yourself. In comparison, sometimes I need to use Fusion 360, and its performance is abysmal. I wish I could do parametric design in blender

I think going from horses to cars is one thing, and going from cars to supercars is a whole other thing. We don't see F1 races and go: "Anyday now I'm going to commute to work in a fraction of the time. Cars can be so fast!". AI is not a categorical technology, it is still a component of the computer revolution, and achieving with it substantially more than we already can today would require the same extreme degree of specialization and constraint narrowing as going from urban driver to racing pilot.

But "code" is still the core substrate of your production, right? And you probably output even more of it. So how is it that you barely code anymore? One could say you're coding way more, and way harder. And you're right: it requires a lot of hope and imagination to conceive further disruptions from this technology. From my POV, it's pretty clear that we have already hit the ceiling. Not because the underlying technology can't improve, but because a technology does not exist in a vacuum, and being one that demands orchestration and supervision, it's clear that it's bottlenecked by its environment. To put it differently, think about how much more could you do with today's version of it, if you were able to read 10x faster, sleep 10x less, and handle 10x more tasks in parallel? So how exactly are we going to create an hyperproductive dystopia if you're already leaving so much performance on the table by simply being a meat bag?

The very question is biased toward the idea that “something is coming”, as if we’re heading to some technological inevitability. I don’t see it happening. This whole “it’s coming” vibe is the narrative attached to the technology, it has nothing to do with the technology itself. It seems like people who buy it are trapped in the fictional reality of a very cheap advertisement, or should I say, propaganda. Your calculator now supports a new operation, this is not armageddon.

Extracting structured data from PDFs is something people have been doing forever, way before LLMs, there are state of the art OCR tools out there you can use as a lib, it’s hardly something that justify using a LLM

Couldn’t all of them be right, though? In the sense that every technology associated with the functioning of the universe had a systemic, spatiotemporal and evolutive dynamic, which appears to be the sensible facets of the universe’s behavior? In that sense computation is no different, just has more primacy. It gives me some Hofstadter vibes, meaning whatever complex thing we produce within the universe is hopelessly bound to be analogous with it.

In the day and age of tweets that move entire markets, countries that have official online profiles and use them as a major mean of public communication, Wikipedia policies feel so lame, mostly because it is not even an encyclopedia, but a website that pretends to be one

It’s also intriguing how an academic article read by no one would pass as valid source but an YouTube channel with multiple videos about the thing reaching millions of people across several countries and languages would not. I understand it’s not about reach, but trust and such. And it’s weird that Wikipedia, a Website, does not trust, in essence, its very medium.

The landing page is the worst place to get in touch with the reality of the product. Take a STEAM store page, it says nothing interesting about the game, doesn’t even has a gameplay video. Much better to just get to know the product with real people with real stories to tell, then the landing page is just the last place you go to acquire it. I never understood people who actually reads marketing copy and goes “hmm interesting”, it’s so vacuous and boring.

Could they have treated subsidies to inference as a sales and marketing expense, though?

What I find curious is that a machine replacing manual work is a new device in the environment, but LLMs are not new a device, they’re a feature on the device, the device still is the computer, and the operator is the same. There’s another conflicting angle, which is that LLMs derive from sampled human work and is paid for by wealth generated by human work, so if you discard human work there’s no wealth to maintain the LLMs nor data to feed it. And I don’t really see how can LLMs produce their own wealth nor how it could succeed producing their own samples. Humans and LLMs are natural complements of each other. I cannot rationally see how this development could lead to a complete breakdown of labor economics without the complete breakdown of the technology itself. Furthermore, human cognition, language, culture, philosophy, science, etc evolve at enormous momentum, with basically zero friction, while training of LLMs is insanely costly and slow. Even if it could self improve, it would still be way slower and increasingly costly (dataset only grows). So advancing this technology without restrictions is the recipe for its own demise, for many possible reasons: no one to pay for it, no material to train it, no supply to scale it, etc. That is, its own potential is tightly tied to the potential to starve itself. So the sustainable thing to do is to find a balance. (This is just to talk about general dynamics, not considering particular social, cultural and environmental effects at each point in the evolution curve)

The issue with tauri is handling big blobs, you either have to serialize them between processes or serve them to the webview via tcp. Albeit dangerous if done wrong, Electron can be configured with direct access to the file system. AFAIK tauri does not have this escape hatch

Get real, kepano. You’re overestimating the consciousness of most casual users. Having godmode, RCE-capable plug-ins behind few safety warnings that most people will happily ignore to get shit done is not good engineering. I understand the constraints. In your shoes I would at minimum make a different version of the app in which you could allow these plug-ins and not put them under trivial banners within the canonical version of the app. You say you have banners, but these sit in the natural flow of the user journey, the options are clearly available and these banners are merely to exempt you from any liability, not to protect the users.

Web stack plus lack of resources to architect the proper interfaces is my guess. These are software written in high level js frameworks, thus using poor dataflow patterns by default, mostly just following what is actually possible instead of employing intentional design, which would require going down some levels of abstraction and maintaining a custom fork of said frameworks. So they probably just architect plug-ins like you would instantiate a library passing a subset of the context the app uses. Basically the simplest workable thing possible. Although the disclosed hack does not mention any particular “vulnerability”. Plug-ins in obsidian are always in god mode, and the alleged hackers just tricked people in using them. Funny how an RCE waiting to happen behind a few popups is ultimately blamed on users. Shame on the developers.

remapping capslock to esc is something nobody whom i've shamed into doing can go back from. it's just night and day. i've been thinking lately that the reason we need hjkl is vim is because the keyboard layout is actually bad for arrows. on typewriters there was no arrows, but on a computer arrows are of primary importance. i think the spacebar doesn't need to be so big, there's no reason for it to be available to both thumbs, and i think moving the small set of arrows into the left or right part of the spacebar position would be so much better for typing because the hjkl hack only work in hacker editors, but we need to use arrows a lot on normal software and it's super bad for your hand if you use it a lot. i started developing inflamations because of the way i fold my thumb to reach for the arrows without moving my entire hand.

Swimming in borrowed or imaginary cash without any hope of paying it back in the foreseeable future.

Who the hell let agents directly use a database? Even humans don’t get this privilege. So, of all things, we forgot how to write APIs now? The article suggests creating a role for the agent directly in the database. What is wrong with you people? The very title of the article defeats its own purpose. They are not designed for this so don’t let them be used like this, ffs.

The same applies to the south. It’s shocking to read tales of people spending hundreds of dollars monthly with coding agents, that’s wholly impossible for the vast majority of devs in South America, even 20 dollars is hard to justify for most households. By economic factors alone, I bet there are a lot more people learning the hard skills in places they can’t afford to be dependent on the tools.

I’m also in this camp. There’s nothing better than to be lost in your own flow. However, I find these moments to be richer when someone is silently tinkering besides you, in sort of a passive interaction. Typical people tend to behave awkwardly when there’s no point or reason in talking while in the company of others. This has to be as much as a deficit as the normative definition of social awkwardness. I could never connect with these kind of people, that are always ruining silence for no reason other than trying to escape their own discomfort.

Like others mentioned, letting the agent touch the code makes learning difficult and induces anxiety. By introducing doubt it actually increases the burden of revision, negating the fast apparent progress. The way I found around this is to use LLMs for designing and auditing, not programming per se. Even more so because it’s terrible at keeping the coding style. Call it skill issue, but I’m happier treating it as a lousy assistant rather than as a dependable peer.

Adobe won’t be hurt by this in the professional market because they have inter-app compatibility and a somewhat consistent language, plus you need their software to work with legacy files. Adobe is cheap, you can get the full suite for a very reasonable price. Competing software is always niche and you need to learn each one individually as they don’t share UX principles nor ontologies. They might be free now, but imagine managing individual subscriptions for each one later on; a nightmare for individuals and companies alike. Just needing to sign-up for multiple apps individually is a headache, all the emails and updates, etc. Unless someone makes a comparable and comprehensive suite, they won’t be actually competing with Adobe.

Typst is unfairly good for doing systematic designs. I wrote a template system for a complex product catalog in a couple days. Then I modeled the clients products list (exported from their ERP) to the schema and generated a hundred pages catalog instantly with flawless layout. Traditional catalog design in InDesign is extremely prone to errors and inconsistencies, not to mention time consuming if done by hand and very brittle if done with the native automation, which does not handle tabular data very well, requiring arcane non-UTF8 encodings. With Typst, if done right and input data is properly treated once, you can wholly skip the review phase which is represents a massive cost reduction. IMO doing this kind of parametric design from a DSL, either for print or digital, is something massively underrated. Surely feels like cheating. Organizing the media files is a bit more time consuming, though, even with automation. But once you organize and standardize the media repo you’re set, as you just need to do the plumbing once.

There’s no conflict here. Using a tool to automate what you have validated to be the trivial parts of a production process is the proper use of the tool. Professional designers also use this bias. For instance, I might recognize that creating a custom font or illustration is not core to my solution, so I can employ an off the shelf font or illustration and focus, say, in the written content. Same principle. The problem is most people won’t even acknowledge or validate the essential aspects of the solution and just iterate mindlessly.

Balancing requirements to achieve something you care about is doing design. I take that by “design” here you mean perhaps a particular interface or media, and you reckon that such element is not critical to your solution. If that’s the case then there’s no conflict at all. By reaching that conclusion you isolated what’s important and are correctly applying energy where it matters. This happens a lot in design, where producing or perfecting media interfaces is not necessary.

The argument is not that only designers can design, nor that everyone should design like a designer. It’s to not confuse shopping for or generating generic solutions with the activity of problem solving. Per Alexander, trivial problems, those that can be solved without balancing interactions between conflicting requirements, are not design problems. So, don’t worry and just pick what you need and like!

On Notes on the Synthesis of Form, Alexander defines design as the rationalization of the forces that define a problem. You’ll won’t find a better definition. But people tend to think design is the synthesis and its results. This misunderstanding of the role of design and the designer is responsible for all the unfit designs we encounter on a daily basis. Anyone equipped with a synthesis tool and feeling empowered to quickly and cheaply generate forms will almost inevitably become blind to the very nature of the underlying problems they set to solve. They’ll be fitting the problem to the available forms. They’ll skip the understanding, the conversations, the conflicts and disagreements, and happily and wrongly assume a design problem can be solved hermetically, in isolation. They’ll think quality is a factor of aesthetics, when in truth, aesthetics is an effect; nevertheless these effects is all they’ll have control over, as it’s all the tool can do. The tool will hinder their ability to be rational; to see the inner structures; to find the hidden but essential semantics; to create the ontologies that’ll support not only the immediate synthesis, but that will sustain the evolution of the design over its lifetime. They’ll be denied the enlightenment that comes with gradual, slow construction; the only place and moment where innovative ideas reveal themselves. They’ll be impoverished and confuse output with agency. I feel sorry for anyone that will think using tools equals doing design, because of the truly marvelous human experiences that they’ll miss, and that could never be replaced by the shallow pride of empty achievement.

Just a thought: The fact that the found kernel vulnerability went decades without a fix says nothing about the sophistication needed to find it. Just that nobody was looking. So it says nothing about the model’s capability. That LLMs can find vulnerabilities is a given and expected, considering they are trained on code. What worries me is the public buying the idea that it could in any way be a comprehensive security solution. Most likely outcome is that they’re as good at hacking as they’re at development: mediocre on average; untrustworthy at scale.