HN user
Destroying unsold clothes is financially the most efficient thing to do. It remains unclear to me how taking actions to maintain higher markups on products would be socially efficient in any way. Companies of course can keep doing it, they just will face financial and legislative repercussions for it.
There's a bunch of lines before the 7th paragraph that say stuff like "suggests" and immediately before the line you quoted, "Although the material is scarce on the subject,". That's just the nature of history, sometimes you just don't have the full documentation even when you're talking about your own company.
Any specific reason as to why you'd want to make that, outside of intentionally enabling fraud?
If it's anywhere close to the same universe as smaller models in its behavior, a lot of time in "thinking" mode is spent on reiterating on any constraints given in a prompt. So the more constraints you give it, the more tokens it will spend going "Hold on, the prompt said I have to dot my i's and cross my t's. Let me go through my work to check that all the i's are dotted."
How is pasting an article into an LLM going to get you a neutral opinion? It'll be at best an 'opinion' that aligns towards the fine tune dataset used by the org that made the LLM. I'd rather people own their own bias and bring something into the conversation rather than act as a mouthpiece for a statistical median.
Conventional web technologies already can handle this kind of use case, either by wrapping an image in a link tag, or using an alt= value.
Sounds like a bad idea in general. Any data use agreements get lost, shadow-IT brews and nobody knows what tools to use, oh and it's against the service terms.
Sorry, how would that work for literally any non-computational science? You just can't submit papers that involve actual reality?
I don't understand why they'd hand over those data points over to the model in the first place. If it's in the context window, it's impacting the output. To ensure that no weight is placed on those factors, they should be sanitizing them out before handing the data over to the model.
I'm imagining it, and it'd at best be the same as if you gave your website to an ad sales agent without any instructions as to what your actual product or target audience is. At worst it'd be writing copy that is fully fabricated, doesn't match your brand language, and opens you up for false advertising claims.
I'd also strongly endorse its on-the-street mobile app counterpart, StreetComplete. Splitting stuff into tiny tasks you can focus on really helps if you're just out on foot in an area.
The post title about it being "pixel-faithful" is a bit strange. I don't see that claim in the repo, and they don't seem to even claim full feature support at the moment. And for the features marked as supported in .pptx's, it does seem that at least slide image backgrounds and bullet point images aren't actually working, and some text objects have inverted text colors. Seems quite far away from being pixel-faithful in fact.