Agreed. This reads more like a very junior dev reads static analysis warnings and extracts it into a constant to satisfy the ide. An LLM would at least give the constant a slightly abstracted name.

> The Chinese lab reportedly needed only two months of training time to reach performance levels that took Google, Microsoft, Meta, and OpenAI years to develop.

Sounds a lot like standing on the backs of giants to me. Why would this blow minds that with newer compute and full hindsight, someone could reproduce something more efficiently?

I feel like I’m missing the point and Google didn’t illuminate any deep article that represented this achievement in novel terms.

So OpenAI doesn’t need to burn through a trillion dollars in the next 10 years after all?

Cool.

I think that if investors see this, that you only need few millions to be almost number one, they will not give them billions so easily. Deepseek changing the narrative of spending money in AI tech.

Right now, they are standing on the backs of giants, but in the future...it could be opposite.

In other countries you are not allowed as a foreigner to buy land near a border. At first, I thought this to be ridiculous but after I put my own aspirations aside, I realized it a tempered approach to request only citizens own the land near borders. It seems logical that we world prevent non citizen owners near military bases or borders for the sake of hampering our advisaries.

I was so sure I could make it give me the right image, but I was wrong! Even demanding the disk be for Apple IIc and/or describing the softer shell wouldn’t work. It gives some strange windows vaporware when I asked to show how to load that (incorrect) disk into an Apple IIc.

Before the social media mega companies we had all sorts of weird conspiracy theory nuts, but it was a lot easier to be skeptical when you only hear about it from the one guy who always has those nutty ideas (we all have that one uncle). But his ramblings can masquerade as credible when they are backed by larger groups people regurgitating the same talking points in unison on social media, filled with ai images and any other tool they can use to appear more credible.

IMO we need to go back to email lists. No more algorithmic feeds controlled by mega companies… for the same reason that our government in the USA suffered as a result of senators no longer moving their whole families to Washington during their term - those of us that are consuming content from feed algorithms are only seeing what triggers them to want to scroll more, not a balanced assortment of info from everyone.

The train has left the station. There's no going back to email lists. What would be better is to build a product that's better than the current ones or force change.

I'm honestly suprised Youtube doesn't get dragged in the media with what they put in that sidebar. They are setting up a generation of adolescents with a bunch of extreme, and rather adult, content that is real-life or masquerades as real.

> The train has left the station. There's no going back to email lists.

Not if it's left up to market forces, but regulation could do it.

> What would be better is to build a product that's better than the current ones or force change.

IMHO, that's a fantasy.

I think I heard this pitched by Neil Diamond 15 years ago. He had an f body or similar old big body car from the early 80 retrofitted with a rotary engine charging the battery… I’m so glad to hear this finally making production.

The author seems to contradict themselves in many places… maybe they just really want it to be true. For example, “allegedly” and “convicted” in this case:

> He was convicted during Tusk’s previous regime (2007-2014) for allegedly abusing his power while pursuing government corruption with “excessive zeal,” but was officially pardoned by then-new President Andrzej Duda in 2015 – a long-standing point of displeasure for the Polish left.

I have a cron job taking a picture of my desktop four times per minute into a dated folder structure. It allows me to zoom back in time (days out weeks back) to figure out what I was doing at a particular time or if I was showing someone a particular bit of code or something like that.

Check out Rewind.ai! I used a similar app (Capturer.app) to do this automatically, and have since switched to Rewind - it's much better for my purposes.

But its constantly using 30% CPU... maybe the screenshot solution is lighter!

Mind posting the script? How much space do the images end up taking per day roughly?

Making this searchable for text would be the killer feature here.

Unless they did business under that name … then they can claim that name in the states where they did business first.

Usually not unless you did business under that name. Squatting a name does little to protect you. If you are selling something under that name first, in a given state in the USA, then you can fight off other people from using that name in that state… applying for a trademark or servicemark helps you extend it nationwide but still requires you to fight off people trying to use your name.