I don’t think “trending” is any better of a criteria for diversification of readings. Especially since the metrics around “what is trending right now” isn’t clear or necessarily objective. More specifically, I think “trending” might be a good venue for seeing what new things are happening now that is of general interest to a lot of people, but I don’t think it’s a good vehicle for getting additional articles from different perspectives.
I do like Tangle for the fact that they grab links to reputable sources from a variety of perspectives, but the selection done more as a human activity far more than AI generated. And I will point to this as anecdotal evidence of the failure of AI to effectively provide a diversification of perspectives on events.
I think the big cause was how much more efficient the Chinese algorithm was vs. the US algorithms resulting in people assuming that Nvidia would sell fewer high price chips.
Some of it what they did seemed like common sense. Instead of using 32-bit numbers, they used 8-bit which dramatically shrunk memory usage. Rather than trying to make the AI all knowing, they made it specialized.
The use of AI by Google is stupid. Last night I was trying to search to see if I could cook pierogies like dumplings and steam them then fry them in the same pan rather than boil then then fry them in seperate pans. I just needed some cooking websites or Reddit discussions. Google gave me an AI summary that wasn’t particularly helpful. Why?
On the other hand, I was impressed using copilot yesterday to try and resolve a problem I was having doing something in R. My Google search wasn’t finding what I was looking for. Copilot answered my question and was able to redo the example a slightly different way when I asked a follow-up. However, the coding problem was due to an unrelated issue with my data frame when I could see that what copilot was doing the same thing I did and it worked for the simple version of my question.
I’m sticking with AI is going to help knowledgable people be more productive, may eliminate some low skill jobs, but I’m still not seeing mass unemployment. It can give you a gazillion different options on how to do something, but most of the time you still need someone to figure out which option is the best based on experience.
I have read in some places that this may not be enough to financially justify the enormous costs of developing and deploying ai as it is currently envisioned. i don’t really know how to even start evaluating an argument like that since it really depends on things like cost of computing which i know little about, but it seems reasonable.
I think this over simplifies what work looks like e.g. a programmer doesn’t just write code. They also have to know what the code is supposed to do and be able to figure out if it’s doing what it’s supposed to be doing properly and efficiently.
Yes, I heard the same, was just throwing out some random thoughts.
I suppose I don’t really understand how Nvidia is worth $3 trillion in the first place. Even supposing a very bright future for AI and a near monopoly for Nvidia.
I’m enjoying my purchase of Nvidia shares yesterday. I’m up more than 6% so far, but I’m sure it’ll be a rollercoaster ride.
It seems like they have unlimited demand for all their stuff right now and for a few years to come. Also, it’s not just AI. It’s crypto and computer games too and maybe the fall of Intel.
I wonder if this feeling about 3 trillion dollar companies is similar to what people thought about the first 3 billion dollar company or the first million dollar athlete.
That would have been either US Steel or Standard Oil. Both of which were monopolies that built the world economy. So is Nvidia the same?
Oddly I do think AI is revolutionary. I’m just not sure how that translates to a dollar value. Which in turn makes me uncomfortable in guessing what news like deepseek does to the going assumptions.
(And I don’t know if athlete salaries are surprising? Presumably a sport makes $X and they pay athletes some % of X.)
Yes. Has been verified many times now. It begins to write a response, and then its text is deleted. The open source model is also censored and either “thinks” about how to avoid criticizing the CCP, or doesn’t “think” and just gives the party line.
He (she?) was carrying bags of something. Could have been clothes. Could have been money. Could have been a newborn daughter on her way to the nearby river.
No matter, as we will have plenty of “stopping tanks” moments in the next four years. Whether someone will be brave enough in this country remains to be seen. The Someone Else Will Do It" culture is thriving here.
Google doesn’t have a problem with it, yet.
Google “tiananmen square tank man”
This is a fun benchmark. Besides being a lot easier to comprehend than math contests and “PhD” thinking, it’s a good reminder that LLMs are still gradually improving.
they do this at football and soccer matches where I live. Tap your card to enter the store, pick out what you want, and it charges you automatically.
They also have an optional facial recognition thing where you can enter the stadium or stores without tickets or card, but that freaks me out a little so I haven’t opted in.
