martopix

12 ms·

91.
by martopix 3y ago
Indeed, and it's a topic that would not be covered by big tech. I think this idea that academia must compete on the same terrain as big AI companies is misguided. There are areas where only academia can be successful, because they woul
92.
by martopix 3y ago
I feel this isn't quite a failure of the Alan Turing or of the UK, but rather of university research vs big tech -- the latter has a significant advantage in this specific field, for a number of reasons. There was an interesting paper
93.
by martopix 3y ago
I'm not sure what you're suggesting. If you want to version control your images you can, and then they will use up size, if you don't want to, don't version control them. In github, you can also just download the master
94.
by martopix 3y ago
It's worth mentioning one of my favorite language facts of all time: the word dram (small unit still used for a portion of whisky in Scotland), drachma (the pre-euro Greek currency), and dirham (the currency of Morocco) are all cognate
95.
by martopix 3y ago
> not a single country should interfere with this I'd rather say "a single country should not interfere with this". IF we have a "simple" internal conflict between local factions, then it should be an internation
96.
by martopix 3y ago
If you end up concluding "the problem is most people", then you are likely missing the point. Rules need to work for everyone. Including those who are newcomers, those who don't have the time to study them, or who aren't
97.
by martopix 3y ago
Populist take aside, this is somewhat true. In Italy there has been talk of electing the president directly, but I think that would be a terrible choice. Currently, he is elected by parliament, and this (surprisingly) has worked mostly well
98.
by martopix 3y ago
It is true that there is a convention not to actually exercise many powers as a King, leaving them to the government. But it wouldn't be an absolute monarchy anyway, in fact it's one of the first examples where the monarch could
99.
by martopix 3y ago
"The boss wants to see you immediately."
100.
by martopix 3y ago
It would be a much better headline if it said "HPV, for which a vaccine exists, causes cancer: get vaccinated".
101.
by martopix 3y ago
The conclusions are something that is well clear to anybody that is familiar with the environment of scientific research. The kind of metrics used to judge research are greatly damaging. I mostly disagree with the distinction made between s
102.
by martopix 3y ago
> The stigmatization of mental conditions in Europe is sky high, to the point that it's barely possible to be seen as sane if you're just seeing a therapist. I think this is just not true. I've never seen so many people ta
103.
by martopix 4y ago
This isn't whining. It's about understanding what their role should be in this context.
104.
by martopix 4y ago
> OpenAI wasn't created to "Do Science". That's exactly OP's point
105.
by martopix 4y ago
The article does not state that "interview questions were regularly approved by a committee". It says that this question was banned because it was leaked to the public. And that he, very reasonably, asked the hiring committee if h
106.
by martopix 4y ago
If you've ever tried machine learning, it's easy to understand what a prejudice is: inference based on learning on biased data. Or sometimes the data doesn't even need to be biased: your stereotyping may be factually true on
107.
by martopix 4y ago
Fantastic for teaching high school physics
108.
by martopix 4y ago
ah right my bad, but the observation stays
109.
by martopix 4y ago
Not only bad people. Everyone has to find ways around it. Even the bureaucrats themselves tell you to ignore the rules and stuff like that.
110.
by martopix 4y ago
These are kibi-meters
111.
by martopix 4y ago
I keep thinking that what I'd really like is a substitute of latex that is based on markdown and CSS. This is a hint of what would be possible, although of course it would take decades to get to the power of latex.
112.
by martopix 4y ago
Correct for the exact opposite reason. Our homes use less and less power as better insulation becomes widespread. Italy for example just had a massive campaign of incentives for people to renovate their homes to better power saving.
113.
by martopix 4y ago
overwork, bad salaries, competition, bad relationship with supervisors are often mentioned but I think this is not the whole picture. I would focus more on research itself. Research can be really shitty. It's extremely open ended, so y
114.
by martopix 4y ago
I feel you. And you know what, it's really ok. It's difficult to accept because sometimes our disciplines have become parts of our identity. But the moment I accepted that I didn't need to love it, and I could move on and do
115.
by martopix 4y ago
Being a professor is much, much more about research than about education. You don't normally become a professor because you're good at teaching, nor because you love teaching (or maybe you do, but not only that). This is a bit sad
116.
by martopix 4y ago
The way I would build a DM with GPT4 is with a few external tools: the character sheets, random dice, and perhaps other things should be available on the side, and fed to the chatbot when needed. Same for a summary of the campaign, so that
117.
by martopix 4y ago
The article clearly states that the "placebo" theory is baseless, so I'm not sure what your "bottom line" that closes with "everything else is a placebo" comes from. What is "everything else" exa
118.
by martopix 4y ago
I asked ChatGPT to be my dungeon master for a tiny text-based dungeons&dragons game. It took a bit of prompt engineering (it would initially believe me if I unilaterally said I won dice rolls and even the whole game). The result was pre
119.
by martopix 4y ago
Then that's what 'interesting life' means to you, which is ok
120.
by martopix 4y ago
I'm in exactly the same situation. Possibly.

More ›