My conspricy theory is that early LLMs have a hard time figuring out the logical relation between sentenses, hence do not generate good transitions between sentences.
I think bullet point might be manually tuned up by the developers, but not inheritly present in the model; because we don’t tend to see bullet points that much in normal human communications.
more like vinegar than spicy vinegar TBH
I want to point out that macaroni and tortilla chip are not even topologically equivalent. Hence, OP’s statement is obviously ridiculous.
Well at least there is no clear conflict of interest /s
Google, despite having a track record of killing services, does have a good track record to keep their phone up-to-date. On the other hand, most other android OEMs stop frequent update after a year or two.
How do you Apache license a LLM? Do they just treat the weights as code?
I hate to disagree but IIRC deepseek is not a open-source model but open-weight?
I have recently applied for a junior computer scientist position in Europe, U.S., and China. The Chinese employer give me an offer immediately, with almost the same salary as Europe and U.S. In addition to that, the Chinese institute promised me a clear path of advancement, which is not included in my Europe and U.S. offers.
My current employer in Europe is a famously well-paying institute, but my salary is only minorly higher than my Chinese offer, in a much more expensive city than Beijing, and I have went through a lot of crap moving to Europe.
The Chinese government is investing a crap ton of money and resource into computing, while the U.S. is denying visa for top talents, I cannot imagine many people will make the same choice as I did, especially when they spent most of their lives in China.
We have these bullshits, and apparently DEI is the problem within scientific research…
LOL, their code is probably written by AI.
I am great, it is the lunar new year and I am celebrating it with my family.
I made vegan fried rice for them. Most asian haven’t had vegan fried rice, because typically fried rice involves lard and egg. However, they loved the vegan fried rice so much that my cousin asked my aunt to recreate it tonight.
I have tuta, bitwarden, firefox relay, and libreoffice, seems to work decently. I am also considering getting a mullvald vpn, but not sure how well it works with torrenting, since they no longer support port-forward.
Oh, yet another speculative execution flaw…
Yeah, rookie mistakes given we already know who is elected.
As if anti-trust law doesn’t exist. It is crazy to me nowadays, most tech startup’s goal from the very start is to sale to a big tech competitor. This certainly should have anti-trust implications.
With enough money, sorry, I mean “campaign fund”, let’s see how things go.
Thank god for Trump to save communism /s
You are right, but I imagine it is harder to sue a company over GDPR if they don’t even exists in Europe…
How do people not think that China is also a hyper capitalistic society, especially in the tech sector. Your data is 100% being sold if you are on any Chinese platform, just like in the U.S.
If anything, Chinese big tech tends to be less privacy-respecting than the west, because they don’t need to operate in area with basic privacy laws, like Europe and California; and there are much less alternative products to choose from because of the GFW.
The founder and CEO of Baidu openly stated that “Chinese people are less sensitive about privacy, which gives us more data to work with” See https://m.163.com/dy/article/DDRTB01Q0511FQO9.html?spss=adap_pc
deleted by creator