jmugan

9 ms·

121.
by jmugan 3y ago
Ah, but it is not one, all of the humans are biased. But maybe the difference is that each human is biased in a different way, whereas the algorithm applies the uniform bias to all. So maybe the humans are like an ensemble method.
122.
by jmugan 3y ago
Even when it doesn't refuse to answer, the paternalistic boilerplate is really patronizing. Look man, I don't need a lecture from you. Just answer the question.
123.
by jmugan 3y ago
I wish they would spend some time working on their spam filter. "My colleague reached out to you last week, but since we are all busy I thought I'd ..."
124.
by jmugan 3y ago
I think of it as "stealing" my attention. Instead of enjoying blue sky as my thoughts wander peacefully, someone takes this from me to yell that I should buy a new widget.
125.
by jmugan 3y ago
To me, it's not so much that reality has a surprising amount of detail, it's that our brains make us feel like whatever model we have is complete. Same thing, of course, but focusing on the brain gets us closer to the cause.
126.
by jmugan 3y ago
Yeah, I paid extra recently to buy higher-tier tickets that allowed me to choose my seats, so my family could sit together. I go to actually choose the seats and then it wants to charge me an additional fee per seat, per leg. There were no
127.
by jmugan 3y ago
until they raise it.
128.
by jmugan 3y ago
When I first saw Interstellar I thought it was unrealistic that people could be brainwashed to forget about space flight. When I watched it again last year it seemed entirely plausible.
129.
by jmugan 3y ago
To me it is more Oryx and Crake by Margaret Atwood.
130.
by jmugan 3y ago
That is a good point that it is a challenge to mix robots and humans. That to me has been the most obvious failure so far.
131.
by jmugan 3y ago
Self-driving cars can communicate more effectively. It's so hard to merge because we can't communicate well with the other drivers.
132.
by jmugan 3y ago
Of course, you need to walk before you run.
133.
by jmugan 3y ago
Human drivers kill 40,000 people a year in the US. It's a catastrophe, but we are all just used to it. Because of that, I'm willing to give self-driving cars some slack. People put cones on them. What would happen if you tried to
134.
by jmugan 3y ago
I remember in 2011 looking into decentralized social networks. I didn't see the point since existing ones worked fine. Time has proven me wrong. The existing ones now don't let me control my feed, and they keep throwing unwanted n
135.
by jmugan 3y ago
I'm not a truck guy, but the Ford Lightning looks pretty cool. I wish it wasn't so expensive.
136.
by jmugan 3y ago
It's worth watching because it is so plausible.
137.
by jmugan 3y ago
Anybody know of something good and simple in Python? I'm more interested in agent-based machine learning in 3d simulation than making games. I'm familiar with Ursina on top of Panda3d. Is there anything new on the horizon?
138.
by jmugan 3y ago
Not only do you pay, you pay taxes on paying (sales tax).
139.
by jmugan 3y ago
I keep thinking that as people lose faith in regular cities they are going to want to live in a Disneyland. Reminds me of Oryx and Crake by Margaret Atwood. Traveling into the general city outside the gates was a big deal.
140.
by jmugan 3y ago
I had the same experience with AppEngine. I kept hitting snags and just couldn't get them solved so I gave up on it and moved to AWS.
141.
by jmugan 3y ago
I moved from Matlab to Python in grad school once my thesis program got too big. It's open, which beats Wolfram. You don't have to manage memory, which beats C/C++. And it has a lot less boilerplate than System.out.println(&#
142.
by jmugan 3y ago
Yeah, I loved Reamde by Neal Stephenson, but I had to put down Termination Shock, which takes place about 100 miles from my house, because I didn't recognize the area at all, even accounting for future climate change. He described Waco
143.
by jmugan 3y ago
It's funny how memory works, "complicated doors" is an almost perfect combination of those two mistakes.
144.
by jmugan 3y ago
Those didn't pop up last time I searched in this area. I knew about Stan but not PyStan. I'll check those out.
145.
by jmugan 3y ago
And I wish they would merge it with the beautiful explanations at https://probmods.org/ . We need a practical probabilistic programming language in Python. We have PyMC, but to use that you have to pull out your old notes on
146.
by jmugan 3y ago
I wish Pyro would do a better job of hiding the implementation details. I shouldn't need to understand variational inference and such just to get the probability of a god dang hot dog. I've tried to use Pyro a few times, but every
147.
by jmugan 3y ago
I agree on the approach, and I try my best, but it isn't easily. You almost have to make food preparation a hobby.
148.
by jmugan 3y ago
I found that you need to dig into the actual code and debug it while it is running to see what is going on to actually use the library (both with Langchain and LlamaIndex). That's unfortunate, but it does show a path to get you to wher
149.
by jmugan 3y ago
True. But they never seem to convey more than "character X is worried about potential event Y."
150.
by jmugan 3y ago
I love dreams, but I hate when dreams are shown in movies or characters describe them in books. I just skim until reality resumes. Dream interludes almost never offer insight and are just boring.

More ›