I found it able to route me to where I need to go if I decide to switch careers and become a train driver, but planning something easy like "take a bus to the train station" is impossible in OsmAnd.
Obligatory reminder that the Dropbox thread ends with "I only hope that I was able to give you a sneak preview of some of the potential criticisms you may receive. Best of luck to you!" The comment didn't dunk on Dropbox as an idea, but pointed out that they would need to highlight their moat wrt copycat competitors in order to convince sceptical investors.
The artist in question is presumably not raising VC money, so concerns about long-term viability of the niche if other artists start imitating the style probably don't apply. (Maybe it's even the reverse situation, where increased production of cracked-glass art raises the profile of the trailblazer and increases the demand for "originals.")
I'd be interested to know why those comments were flagged actually. They don't scream AI and no-one has replied calling them out as AI, etc. But the vast majority are dead.
That's why. Boring, bland, etc. That account's M.O. is basically "write a paragraph that says nothing." Fwiw, I do think AI can be indistinguishable from dumb, boring people, but usually those kinds of people won't be on HN.
The account was immediately shadowbanned after re-awakening from a long period of inactivity.
I agree it doesn't seem obviously AI. The early comments are all in the same writing style and smell human. Lots of strong opinions e.g.
"logged in after years away and had basically the same experience. the feed is just AI slop and engagement bait now, none of it from people I actually followed." [about Facebook]
HN has got a big problem with silently shadowbanning accounts for no obvious reason. Whether it's an attempt to fight bots gone wrong or something else isn't clear. By the very nature of shadowbanning there is no feedback loop that can correct mistake.
Pretty sure they weren't shadowbanned immediately, since people replied to some of those [dead] comments. Most likely the shadowban was applied retroactively after posting the more obviously generated stuff.
The problem is not the use of blackletter, but the narrow spacing and copious use of abbreviations to cram the text into two rectangular columns. That was certainly not an unusual goal to have, and you can see the same in handwritten manuscripts https://commons.wikimedia.org/wiki/File:Bible.malmesbury.arp... though executed less perfectly, but that doesn't mean it's not form-over-function.
I gather that form-over-function is not as binary as I imagine. That and I’m at fault for assuming that because that style of typography was prevalent people enjoyed reading it or found it legible.
I doubt that a regulator would be satisfied by the kinds of explanations this provides and the interventions it enables.
Suppose somebody put an LLM in charge of an industrial control system and it increased the temperature so much that it caused an accident. The input feature attribution analysis shows that the model was strongly influenced by the tokens describing the temperature control mechanism, concept attribution shows that it output tokens related to temperature, industrial processes and LLM tool-call syntax.
The operator proposes to fix this by rewriting the description and downweighting the temperature concept in the output, and a simulation shows that with these changes the model doesn't make the same decisions in this situation anymore. Should the regulator accept this explanation as sufficient to establish that the system is now safe?
If the controller has just a few parameters and responds approximately linearly to changes in its inputs, you can in principle guarantee that it'll stay within a safe zone. But LLMs have a huge number of parameters and by design highly nonlinear behavior. A simple explanation is unlikely to reflect model behavior accurately enough that you can trust its predictions to hold in arbitrary situations.
reply