It inspires me to tackle a project I've been holding off on for many years: OCR my grandmother/great-grandmother's cookbook. It's about 100 pages of collected and annotated recipes from the 1930-1980s.
OCR and AI have become sufficiently capable (as you've demonstrated) to properly scan, index, and classify the recipes into something I can share with relatives online or as an ebook.
This is what I've been doing for a couple years now: having AI help to code/test projects that I've had in my long TODO list but would never realistically started/completed. AI is now pretty capable of producing decent code if your specifications are decent.
I still think that non-programmers are going to have a tough time with vibe coding. Nuances and nomenclature in the language you are targeting and programming design principles in general help in actually getting AI to build something useful.
A simple example is knowing to tell AI that a window should be 'modal' or that null values should default to xyz.
I completed three "micro apps" solely for myself (well, one for my wife and I) over the last month, all vibe coded. I sent two of them (sort of micro mindfulness tools) to a therapist friend. "WOW, I didn't know you could do all that!"
What I told her was, I, in fact, could not do all that. I'm a marketer who's been in software companies for 15 years. Zero coding ability. However, through repeated exposure to programmers, PMs, and designers - and because a major part of my role is to develop a nuanced understanding of how our products create value for users - I've learned how to think like a product designer and developer (albeit to a non-professional degree).
That's the part of the "anyone can now build software" sentiments that seems to be skipped over. Anyone can now have their ideas coded for them, but that doesn't automatically create the ability to make all the right decisions during the design and build process. It doesn't create thoughtfulness and consideration and the willingness to say "no" to a feature because, while it seems cool, it will detract from the core purpose of the overall tool.
I've a similar history of AI use as you but every so often I simply describe what I want and try out what it creates. Honestly in these past 2 years the pace of improvement has been stunning.
Yesterday I was inside one of the tools that a just build it prompt created and I asked it to use a NewType pattern for some of the internals.
It wasn't until I was in bed that I thought why? If I'm never reading that code and the agent doesn't benefit from it? Why am I dragging my cognitive baggage into the code base?
Would a future lay vibe coder care what a New type pattern is? Why it helps? Who it helps?
I think the pedagogy of programming will change so that effective prompting will be more accessible.
I do think about this a lot. On the one hand, you might be right, and it may not matter at all. On the other, we often do that kind of stuff because it makes it harder to slip up by accident and/or makes the code easier to read and understand. These things are surely helpful to AI agents in the same way that they are useful to people?
I guess it depends on whether the extra time you could invest in that kind of thing pays back in terms of context windows, code quality or speed of AI code generation.
"Ok, here's a static bar on the top of the page" as it disappears as you scroll. You can now no longer click anything else on the page. You said "can't click away" and it did show up on top. As a 30 year coder that never did any UI, this is this shit I run into constantly. I can create awesomely cool and fast back end stuff, and can create better UIs than I've ever been able to, but not knowing the nomenclature trips me up constantly.
In pretty much all cases, the companies in question had peaked were experiencing declining growth and attempting to do a hail-Mary... and failed miserably.
Compare Digg and slash. One completely died, the other has stuck with its formula and hasn't disappeared, but has just faded into irrelevance.
To protect my privacy, I have a photoshopped drivers license with a photo of my dog that I've successfully used for verification (e.g. AirBnB) in the past.
Though, with AI being used I suspect it wouldn't pass any longer.
Huh. Can you do that? I wonder what is legal status of this. I used to make all sorts of fake IDs (pretty good ones!) when I was a teen (you know, for purposes such as going to clubs, buying alcohol), but of course this is literally a crime, and not even a "minor" one. Apparently, back then it didn't bother me much, but with age I became more cowardly, I must admit. So now I use my passport data more often than not, even though I am not really a fan of the idea of giving a scan of your documents to some random guy on AirBnB (although, with some obvious caption photoshopped on top, to make the scan less re-usable). I mean, it's just a matter of fact that everyone requires them, and it also has that weird status of "semi-secret thing" that you are somehow aren't supposed to give to anyone, and I still have close to zero understanding of how that works.
So, I suppose you shouldn't give your fake id (digital or physical) to a government officials. It also seems "obvious" that it's similarly unwise to give it to a bank. But you can do that to a random guy on AirBnB? A hotel? To a delivery service (Uber/Wolt/whatever)? Dicsord? Where is the line between a bank (a private commercial corporation) and Discord (a private commercial corporation)?
>>But you can do that to a random guy on AirBnB? A hotel? To a delivery service (Uber/Wolt/whatever)?
The "legal" line is usually around fraud - trying to obtain some financial gain by providing false information. There is nothing to gain by giving a fake ID to discord - but it probably violates some rules around unathorized access to computer systems.
I guess I assumed it’s illegal in that you are using an image to tell a lie in a transaction… like any other kind of forgery - but what i’m actually unsure of is posessing a jpg of an altered drivers license illegal? Seems different than a physical license.
I was referring to the concept of "ceci n'est pas une pipe", and that even just digital forgery of an ID can constitute a crime that can be prosecuted independently from anybody suing.
Of course I highly doubt they'd sue. They either just don't let you in or you abandon them. I'm with the latter.
I’m not a lawyer, but i’d guess that possessing a jpg of a fake id is treated differently under the law than a physical forged id. Once you use it to defraud someone, that’s probably treated the same, but just owning the jpg?
Yeah I agree. There is always some risk about government ID. Long gone the day that ppl could forge one relatively easily, when ID was just a piece of well made paper.
At least where I'm from, the forgery or the possession of a forged ID is a criminal offense in itself, not matter the intent or whether it's actually used.
I'm not sure that photoshopping a dog in place of the portrait would qualify to though. It's immediately obvious that it is neither you nor a valid government issued document so doesn't that preclude it qualifying as forgery?
Yes, it is indeed not always clear what constitutes forgery (Germany).
> A document in the classic sense requires an embodied declaration of intent that identifies an issuer and is suitable for providing proof in legal transactions. In the case of a lawyer's letter, the signature is an essential part of the standard repertoire of authenticity.
So removing some parts to make it _could_ make it safe, to Not create a "risk of confusion":
> Even if computer processing creates the appearance of a genuine document, the typical characteristics of the original must be present to establish a serious risk of confusion. Likewise, the BayObLG did not consider the offense of forging evidential data according to Section 269 of the German Criminal Code (StGB) to be fulfilled.
Off topic, but I love how every country has its weird abbreviations that seem obvious but really aren’t, like BayObLG for Bayerisches Oberlandesgericht (Bavarian State Superior Court) or something close to that. Or how every British cop show assumes its audience knows exactly what a DCI is, as in “This is DCI Foxwaddle and I’m DCI Rugby-Botherington, may we have a word?”
Obviously you can go further: What if you just draw up the whole thing with a pencil? What if it's an ID identifying you as a citizen of Nowhereistan? Where does freedom of artistic expression end and a forgery start?
Again, IANAL, but I suppose it would qualify at least as fraud once you try to use even an obvious fake as "proof of identity" in place of a requested government issued document. In that case, there's an intent to deceive that's hard to deny, even if it's just about age verification for access to a digital platform.
I could imagine in court it might come down to details, like whether it's sufficiently similar to a real ID at a glance, or whether tamper-proof marks of an official ID were copied as well.
In any case I wouldn't want to risk up to a year of jail time over a joke.
Youtube flagged one of my accounts as a teenager because I watched a few pop videos (lol) and I was not able to trick it with fake IDs, though I didn't try all that hard.
I've been grabbing music from youtube for years. I don't mean commercial music. I mean talented enthusiast who does not sell their music anywhere. Rest assured, it will absolutely be gone one day, and they way things are going, it feels like it will be sooner rather than later.
I tried to do this when LinkedIn forced me to upload an ID. It didn't work unfortunately. I see the good in this but I know it will be abused. I want to run away but I don't foresee any way that the powers-that-be will let the common person use the Internet without an approved ID in the future.
Well then what was the point? If you gave them an ID that matches your name and DOB, they still got an identity vector that can conclusively match to your physical, government-acknowledged identity.
Not having a correct photo or license number didn't really mean anything to them if their OCR didn't have any half-decent verification that would look at the fields where that information was expected to be, anyway.
There was a story a bit ago about people using video of someone turning their head from side to side to trick these systems. And of course naturally people will easily get past it.
An interesting approach to the worsening signal-to-noise ratio OSS projects are experiencing.
However, it's not hard to envision a future where the exact opposite will be occur: a few key AI tools/models will become specialized and better at coding/testing in various platforms than humans and they will ignore or de-prioritize our input.
Lord. I can see this quickly extending even further into HR e.g. performance reviews: employee must 'speak' to an HR avatar about their performance in the last quarter. the AI will then summarize the discussion for the manager and give them coaching tips.
It sounds valuable and efficient but the slippery slope is all but certain.
Might be. For highly sensitive messages, Russia still uses physical delivery of typewritten letters. This is because they (rightly) distrust digital security models.
Since the start of the war, Russia will rather admit incompetence (a soldier was smoking next to an ammo depot) than admit Ukraine succeeded in a military objective.
Absolutely gorgeous and extremely smart creatures. They generally bond with one care giver and when that person dies it really is a traumatic event for the bird.
Birds that are traumatized pick out their feathers and look terrible. You can tell from the videos of their birds that they are well looked after.
This organization (and those like it) are fantastic!
It inspires me to tackle a project I've been holding off on for many years: OCR my grandmother/great-grandmother's cookbook. It's about 100 pages of collected and annotated recipes from the 1930-1980s.
OCR and AI have become sufficiently capable (as you've demonstrated) to properly scan, index, and classify the recipes into something I can share with relatives online or as an ebook.
reply