swframe2

9 ms·

61.
by swframe2 4y ago
1. Focus on positive thoughts and suppress the negatives ones. 2. Never stop learning, exploring, challenging yourself; the goal is to live an interesting life. 3. Pay attention to details; try to be prefect but accept that is not possible.
62.
by swframe2 4y ago
DreamFusion can learn 3d shapes without a 3d dataset. Next step is to build a model to learn 3d motion from videos. This should get us closer to mastering language.
63.
by swframe2 4y ago
I turned down offers from good friends/classmates who were VPs at early start ups (viaweb, tripadvisor, paypal, vmware, google, yahoo).
64.
by swframe2 4y ago
Need an open source project that tries to build it. (There is a lot of hard problems to solve but it is worth the effort) Need to start with stick figures in 2d worlds. Next upgrade to voxel 3d worlds. Finally upgrade to 3d worlds rendered
65.
by swframe2 4y ago
We need to build a lot more renewable energy projects and start converting sea water.
66.
by swframe2 4y ago
The large language models are very good at prediction. We are making good progress getting them to reason. We need to figure out how to get them to use an internal "blackboard".
67.
by swframe2 4y ago
Check out this video https://www.youtube.com/watch?v=HoKDTa5jHvg of the math. A lot of other sources skip over the math to give you the big picture. After that you will find the video very helpful in understanding the math.
68.
by swframe2 4y ago
Sounds like we need a "github copilot" for the shell. bash on linux can already tell you which command you need to install if necessary. I really don't want to keep cheatsheets any more. I just want to write a shell comment a
69.
by swframe2 4y ago
We are already doing large-scale climate geoengineering; just not with a goal of making the world livable. There are too many people who want to believe lies. We need natural selection to rescue our planet.
70.
by swframe2 4y ago
The math used in ML papers is very diverse. There is too much to learn. It is easier if you pick a problem and learn the math for it. Find a group that is already working on that problem and ask them for best way to learn the math for it.
71.
by swframe2 4y ago
See this playlist on semidefinite programming: https://www.youtube.com/playlist?list=PLqwozWPBo-FtOsjrla4jk... for the surprising way factorization can be used to partially solve a NP problem.
72.
by swframe2 4y ago
I think it is better to focus on getting kids to do creative projects. For example, parents should get them to write and illustrate a story once a month, or learn to play a music instrument, etc. If kids are doing interesting projects that
73.
by swframe2 4y ago
I wonder if there is a more important paradox "wealthy people trying to increase their wealth will influence government laws so extremely selfishly that the world will suffer a catastrophic event that drastically lowers human populatio
74.
by swframe2 4y ago
I would change this to. Find a project you would love to do (software, hardware, art; anything you love to be really good at). Make sure other people would like to see that project (ask your friends and see what others have done). Then crea
75.
by swframe2 4y ago
Check out https://zivadynamics.com/ . It has the realism but the models need to be manually posed. Using the low-poly from efforts like this one, Unity will make it easier to pose the high res Ziva model. There are several
76.
by swframe2 4y ago
This will eventually link up with MetaHuman, DeepMotion and ZivaDynamics to get the high res end result. There a lot of money to be made. It is will happen in the next few years. We're very close.
77.
by swframe2 4y ago
I think the problem with the news is that it focuses only on the problems; They should do some research so they can focus on presenting the solutions. I think I would subscribe to a channel that tells me how to improve the world's prob
78.
by swframe2 5y ago
I don't think companies buying up homes are acting in the public's best interests. I think we should require companies like zillow to build the homes they sell. They shouldn't be allowed to buy existing homes unless it is to
79.
by swframe2 5y ago
I got stuck in a game. My 12 year-old offered to help. Within a few minutes he solved several problems that had blocked me for days. It reminds me of the intel tik/tok product cycle. Look inside for a while then look outside for a whil
80.
by swframe2 5y ago
Here is one that shows how to implement Transforms using excel. https://www.youtube.com/watch?v=S9eKuRVigjY
81.
by swframe2 5y ago
I wrote several scripts that temporarily add call tracing to the source code. The scripts take a few days to write but I've used them on many projects. Note there are code parsing libraries that can help you. For example: def foo(args)
82.
by swframe2 5y ago
I have seen 1:1s used more strategically by employees who have 1on1 with managers outside their team. These employees often get promoted quickly. Note, at performance review time, employees are stacked against each other. If the other manag
83.
by swframe2 5y ago
1) don't learn from your mistakes; learn not to make mistakes. Look at what others have done; there are a lot of blogs about what when wrong with similar products; make sure you don't repeat other people's mistakes. Take a ha
84.
by swframe2 5y ago
When Dartmouth accepted women they increased the class size so they still admit the same number of men. It seems like a good compromise. I'm sure it is not perfect but I think it should be better than holding the class size fixed. (Not
85.
by swframe2 5y ago
I usually start with a program that visualizes the problem. Then I repeatedly implement small steps towards the solution and confirm that their effects are as I expected.
86.
by swframe2 5y ago
You posit that "the majority of liberals-leaning American"s hold the same beliefs in the "America-divided narrative" and yet conclude that everyone is "surprisingly united ideologically". That seems like a cont
87.
by swframe2 5y ago
To paraphrase Hanlon's razor "Don't confuse reluctance with incompetence". The US is a rat's nest of different interests; it is too internally divided to fix hard problems. There are always wealthy political donors
88.
by swframe2 5y ago
When tensorflow was the dominate api, fast.ai made a lot of sense. Nowadays pytorch seems to have won and it is easy to use directly and there is PyTorch Lightning if you want things even easier to write. There are a several videos on youtu
89.
by swframe2 5y ago
I started sending myself email. I have a filter which moves them to a dedicated folder. I sometimes post the content of the webpage into the body of the email if I worry the site might go away. I will reply to my messages when I want to sav
90.
by swframe2 5y ago
In general, getting a phd is the best way to go but it is not the only way. "The AI Epiphany" channel by Aleksa Gordić is worth watching. Check out his origin story: https://www.youtube.com/watch?v=SgaN-4po_cA He

More ›