You know what is the funniest thing. I used to drink all these different Ales and artisan beers.
After I quit drinking alcohol. I used to have a 4 pack of these 0% Ales/Larger that was supposed to taste similar. The packaging differs from the regular beer in that it has a silver top instead of dark blue on this particular. I picked up the alcoholic ones inadvertently as they were in the wrong place on the shelf.
When I took a sip, I thought it had gone off. It tasted terrible, like poison! Obviously once I checked the can, I realised my mistake. I gave them to my rest of the pack to one of my neighbours I think.
> Nobody has ever walked past a photograph because they can't inspect its digital authenticity hash
That the average person hasn't thought about this doesn't mean it couldn't become a thing in the future. People do value authenticity and genuine things, though I agree the particulars aren't relevant in a lot of cases.
This is a (very expensive!) toy camera, but I could see traditional camera companies like Fujifilm, Canon, etc, incorporating this tech later down the line.
Good UX means including translations for supported languages, not telling the user "do it yourself by highlighting content".
Not translating entire articles to a language you don't support has the easy remedy of letting people select the text and use third party tools to support their specific use-cases. But not including translations for your clickable content for languages that aren't supported are the literal practical limits of ability. I would rather my apps work for people in languages I do support, with full accessibility (and minimal scripting overhead), than to have them work poorly for keyboard-only users in all languages, regardless of my app's support for them.
Again, we're talking about the stuff that should be iconic. Things that can literally be represented by icons. Buttons and tab headings. Things that you shouldn't actually need translated AT ALL, much less into every single language there is.
Even when the language is supported I have had GDPR popups block that language selection. The text in the popup was also not selectable which made it very hard to read what I was or was not agreeing to.
I know you're not asking me, but I would really love the "copy" feature added to ALL context menus.
Right clicking a standard anchor element gives you the "copy link" option, but you don't get to copy the word without having it selected. Would be nice to just have a "copy word" feature, for starters. Could even be expanded so that it auto-selects the text after copying it so that if you wanted to copy more than just one word, you could expand the highlight (with the little widgets on mobile, or with keyboard/mouse selection in that one state on desktop) and then get a "copy text" option that copies all of the selected content.
Zim doesn't have a native phone app and syncing, though, and that's a big draw of Obsidian. It's plenty secure if you don't install plugins all willy-nilly.
Of course it depends on the context, but firing someone over one offense of calling someone a dickhead is quite reactionary and punitive. Grow up. Sometimes people say things they regret in the moment, it doesn't mean someone and potentially their family should be affected. This is why you have policies, writeups and whatnot. Not jumping at the chance to fire someone.
These supposed "productivity gains" are only touted by the ones selling the product, i.e. the ones who stand to benefit from adoption. There is no standard way to measure productivity since it's subjective. It's far more likely that companies will use whatever scapegoat they can to fire people with as little blowback as possible, especially as the other commenter noted, people were getting hired like crazy.
Each one of the roles you listed above is only passable with AI at a superficial glance. For example, anyone who actually reads literature other than self-help and pop culture books from airport kiosks knows that AI is terrible at longer prose. The output is inconsistent because current AI does not understand context, at all. And this is not getting into the service costs, the environmental costs, and the outright intellectual theft in order to make things like illustrations even passable.
> These supposed "productivity gains" are only touted by the ones selling the product (...)
I literally pasted an announcement from the CEO of a major corporation warning they are going to decimate their workforce due to the adoption of AI.
The CEO literally made the following announcement:
> "As we roll out more generative AI and agents, it should change the way our work is done," Jassy wrote. "We will need fewer people doing some of the jobs that are being done today, and more people doing other types of jobs."
This is not about selling a product. This is about how they are adopting AI to reduce headcount.
The CEO is marketing to the company’s shareholders. This is marketing. A CEO will say anything to sell the idea of their company to other people. Believe it or not, there is money to be made from increased share prices.
Many people use tiling window managers, so title bars become useless. The first thing I did in configuring ghostty was disabling window decoration for this reason.
Let's be honest, it kinda sucks. The commands are barely explained it feels more left as an exercise for the reader to do the puzzle solving of whats trying to be communicated.
Honestly if these processing chain diagrams just had a rollover where if you roll over parts of the command or the block in the chain and the other part highlighted with a description of what the switch was actually doing then a lot more people would be able to understand this, especially if real world before and after examples of outputs were included.
Instead it's <diagram of the chain> <raw string of the command> "Note: one caveat about something"
I am sure there exist people who live and breath media/codecs and they're reasonably fluent at getting ffmpeg to do what they want because of a tremendous amount of practice.
But for the vast majority of folks who only occasionally use ffmpeg to do something, the complexity of it is so outrageous it feels like a parody. Literally (I mean literally) THOUSANDS of options/flags. It's just too much for a human to navigate. Of course we're going to "cheat" or just google up something similar to what we want. If an LLM can handle it, even better.
I sympathise with the overwhelming sensation of the ffmpeg command line arguments.
But the more you familiarize yourself with a/v streaming and transcoding, you soon realize why you need such amount of control.
I mean, with ffmpeg I can easily combine 3 audio clips, 5 subtitles and a separate video, cut away first 25 seconds and the last 5 minutes of the resulting clip, resize it and change the aspect ratio, reduce audio to mono and specify output codecs for audio and video.
And this is still a pretty simple example of what one could want to do.
Ffmpeg has countless other amazing features, demanding more arguments.
How about for example camera stabilization? (-vf deshake)
How would one even start to explain all of this to an app without thousands of command line arguments?
The whole subject is incredibly complex and ffmpeg is by far the most amazing project in this space.
Without ffmpeg, there would be no youtube in 2005, no plex at all and really the whole of modern social web would probably have happened later if not Fabrice was such a fantastic guy :-)
And still I cannot create .ts files (with mp4 inside) using FFmpeg that will be accepted by my dad's TV media player. I have to put them through Avidemux, because somehow it uses a better TS muxer. More compatible with the TV.
