@samokorosec
Product designer living and working in Vienna. Kicked PDF SDK UI standards up a notch, improved regulated software development tools, helped friends create Audio Plugins and Synths and drew hundreds and hundreds of icons in the process.
As a public servant, I write weekly notes to let people see what I do.
I attended various kick-offs, including for new guidance, new websites, a support hotline, a blog post, and a white paper.
I participated in another #ServiceStandard intro workshop and attended events on AI and digitalisation.
Probably both. 👍
Why would people do it? Why pay for and give away their business data to a vibe-coded “startup”, when doing it in-house at least means one doesn’t have to deal with the risk of the startup disappearing overnight?
It all sounds pretty fucking idiotic and desperate.
It’s absolutely amazing to watch all the “AI can help me create a business in a week!” posts, as if no one will ever ask “If it can be done in a week with AI, why would be pay you in the first place?”
It’s all just so incredibly stupid. 😞
OMG WHAT DID YOU DOOOO?? 😱
Open UI has been working on speccing out the focusgroup attribute: a declarative way to support roving focus & add keyboard nav to composite widgets like toolbars/menus
We'd love your feedback!
Learn more & see open questions: developer.chrome.com/blog/focusgr...
H/t to Edge folks for prototyping
A technology being a “huge positive contributor to society”. is an absolute statement. Saying “the negative effects should be discussed on a case by case basis” when that statement is questioned sounds really weird.
Dunno — does that potential positive outweight the potential negatives of people trusting LLMs too much when it comes to their health, legal trouble, and general knowledge retrieval? Does it outweigh students not learning how to do research and formulate thoughts?
Looks really nice 👌
I wonder how data poisoning will play into this over time. It would be _really_ surprising if not at least the biggest players are already researching the shaping of the data/information landscape that’s needed to mess with fully automated AI decision-making systems.
also some excellent examples amongst the replies
( interactive applications that illusrate statistical concepts, that is)
"Surely one glance at the OpenAI API" - that’s the mistake in the assumption right there. 😶
I saw a Tumblr post that was like “Instead of declaring my day will be bad because I slept like shit, I now say that I am open to it improving as it goes on” and it has changed my general daily outlook
Just published a small enhancement and bugfix release for the #Webmention for #Craft CMS plugin.
It adds two important details though: lazy loading and alt text support for avatar images. 😎
github.com/matthiasott/...
The problem with the "You can vibe code a SaaS now in a week!" story, and why those "startups" don’t have a good chance at surviving regardless of tool pricing, is that the second part to the story is "And so can everyone else!"
IMO this is why many react negatively to it. It’s a false promise.
In truth we should start differentiating between criticizing LLMs and criticizing AI companies and their promises and behaviors, because those are two separate debates that are getting mixed up.
The way social media works, I do not have high hopes for the debate to calm down anytime soon, though.
So I’d go with Cynical Hucksters egged on by the Stock Market™.
The big problem I have, personally, is that I think that such behavior is actually dangerous for the tech; oversell and overhype and you just might set up the tech to "fail", as its users will reap constant disappointment from it.
I’ve seen a couple of small-time CEOs hop on the AI train for that sweet, sweet investment money, and those were people very aware of the tech’s limitations whose investor slides implied near-magical attributes (in terms of being a realistic development in the next ~5 years) to "AI Agents".
5. There are criticisms of LLMs that are much more plausible. They are built on a large scale expropriation of the intellectual commons. Demonstrably true! They involve a radical and extremely worrying derangement of power relations in favor of a tiny number of men with weird beliefs. Also true imo.
Absolutely; I’ve read someone comparing the tools to the arrival of Desktop Publishing in the 90s, and I tend to agree.
The technology will democratize the entry to software development a great deal, and the baseline that defines a software developer will rise. Overall good, if it stays affordable.
Or, in other words:
The tools will not magically fix shitty developers who don’t give a rat’s ass about security or don’t understand architecture.
Just as the tools can amplify the best developers’ quality output, so they can amplify incompetent developers’ shitty output.
The problem that many have is that the marketing and hype are overselling the tools by a whole fucking lot here, and a big part of that is that the whole industry is selling a promise of "This will make you absolutely competent" when it’s actually more of a "This will multiply your competence".
The whole LLM debate suffers from a complete and total lack of nuance. I’m afraid that "It’s just deranged, the best coders I know…" does not help here, either.
It’s a tool with very real limitation, and very real benefits. People who actually understand the tool will benefit greatly, yes.
Bubba is Clinton, or?
Absolutely. It’s the kind of “investment” I’d expect from a game developer in the 90ies do because “fonts” were done by Bob the Game Designer, bitmaps and rendering code both, in less than a week because It Just Isn’t That Important.
But 2020s on a product whose main feature is “reading”? Yuck.
Sounds tedious tho… 🤔
(And even 15+ years ago, those arguments seemed like a post-facto explanation by companies who just didn’t want to invest into it, since macOS could render fonts and hundreds of pages of PDFs decently.)
A ton of shitty typography that I’ve seen and read about on devices comes from font rendering libraries that skip hinting and anti-aliasing and what not for “performance reasons” and fonts being designed with a Windows-ish idea of “fitting into pixels” — ideas and argument from 15+ years ago.
From what I’ve seen and read, the side effect if building consensus a la Europe is the sluggishness and a somewhat opaque process. No Great Leader™ to loudly announce policies and push them through quick.
But there is way more movement in Brussles for at least a year now.