I feel like everyone knows it’s a bubble but that people are expecting it to be like the Dot-com bubble where a small amount of companies succeed, hoping that they invest in those companies.

Thanks

Could you link to the post / bot account this is about?

Network effect / discoverability and free CI time

True, I understand why this is better than all those companies scraping it individually (for both ability to opt out and site load), but the way they handled opt out is still quite silly.

I do wonder if this is similar to Mythos, where the company deliberately stokes fears to get people talking about it and spread the assumption that it will actually happen and is a big deal (and to avoid people’s first question about whether it will actually work).

Luckily it probably won’t get made and the company will just rugpull their investors

Better than that

Google translate said Breton

What about the dopper?

This reminds me of something I sometimes see in shows on like Netflix and other media. I can’t remember a specific example, but you often have generic anti-capitalist comments from characters (often portrayed as edgy). It often feels a bit, artificial, like a “fellow kids” moment but politically, I guess? Like activism as a prop / character trait, inserted into a multi-million media production. Maybe someone else can better put it to words, if I had more time I would’ve written a shorter comment

Deepseek recently published a paper in which they describe that vision tokens contain more information than text tokens and that this can be used to compress context.

We present DeepSeek-OCR as an initial investigation into the feasibility of compressing long contexts via optical 2D mapping.

Experiments show that when the number of text tokens is within 10 times that of vision tokens (i.e., a compression ratio < 10×), the model can achieve decoding (OCR) precision of 97%. Even at a compression ratio of 20×, the OCR accuracy still remains at about 60%. This shows considerable promise for research areas such as historical long-context compression and memory forgetting mechanisms in LLMs.

It reminds me of LLM caveman speak, it used to have another option to use Chinese instead of English. A language like Chinese is seemingly better at encoding information in fewer tokens and I think this is the same mechanism why OCR tokens work so well.

That said, I also doubt that voice messages are more efficient than text prompts, but it’s best not to waste too much time engaging with these sorts of LinkedIn posts (and LinkedIn in general).

Just woke up. It’s 6 am and 28 c.

You have been blocked

You are unable to access this website

Is the website also only available in Canada?

It’s been a while, but I think it’s based on a trick to add comments to Markdown. I tried it with Tesseract, and none of the lines rendered, but only if I add an empty line in front.

It’s a bit of a hack, but it can use existing endpoints and only needs some minor changes on the client side. The alternative would be implementing support for it in Lemmy, Mbin, PieFed, etc. but that probably takes a while and can also be done later when it has proven its usefulness.

Edit: I think I found the explanation behind the syntax https://stackoverflow.com/questions/4823468/comments-in-markdown

It was an image from this GitHub issue, but it seems GitHub now adds a JWT to image links that expires?

I know, I opened an issue on Voyager’s GitHub last year to suggest changes

Some communities don’t have specific rules, and some have several. It can be quite hard to judge why something is breaking rules. On the video’s community for example people often report videos, so I have to go through the entire thing to figure out what’s wrong with it (and some people post videos that are more than an hour). Sometimes content is reported because of something the people who created it did, something that is not apparent when watching the video.

I’d say go ahead but make sure it produces accurate enough results and make sure to add something like [AI Transcribed] in front so people can take the potential for additional errors into consideration when reading it.

Also, if you’re using an online service make sure you’re using something that doesn’t use it as training data. Many (probably almost all) artists / photographers won’t appreciate that.

Almost all OCR tools use machine learning AFAIK, the commonly used Tesseract OCR software also uses a neural network.

It certainly isn’t AGI, but AI just means machine learning nowadays.