alexcnwy

26 ms·

181.
by alexcnwy 8y ago
Do it if you are so passionate about the topic that you’re teaching yourself anyway but need to set aside a chunk of time to really get the momentum you want. Don’t do it if you’re hoping it will pull you out of a rut (there will be lots of
182.
by alexcnwy 8y ago
I found a combination of these Ayurvedic herbal supplements taken in the morning before food to be beneficial (do your own research / I am not a doctor / etc): 1. Ashwagandha 2. Gotu Kola 3. Bacopa I’d also recommend considering
183.
by alexcnwy 8y ago
For example good luck getting mayonnaise with your fries at any of the fast food chains here. They have mayonnaise and they put it on burgers but they “CANNOT” give (or sell) it to you in a separate container for your fries.
184.
by alexcnwy 8y ago
I’m in Tokyo right now for work (my startup NumberBoost won an innovation competition with NTT Japan) and it’s so wild how simultaneously forward and backward things are here. On the one hand, there are so many things they do here that make
185.
by alexcnwy 8y ago
I’m in Tokyo for work and it is surprising how many flip phones there still are here. Not sure if they work in the US but might be worth looking into shipping from Japan.
186.
by alexcnwy 8y ago
It’s more tax efficient too. Genuinely not sure why people are downvoting you.
187.
by alexcnwy 8y ago
I build a lot of computer vision systems and did my MSc on classifying what’s happening in video and I am VERY skeptical that it’s possible to prevent this from happening (without having a person watch every video). You can edit a video to
188.
by alexcnwy 8y ago
I agree it could cause harm but I’m interested in what harm you think such technologies could cause?
189.
by alexcnwy 8y ago
This brings up a very interesting and important question around differential privacy. If I scrape millions of photos from Facebook (including yours) then train a differentially private model that can extract features from a new face, is tha
190.
by alexcnwy 8y ago
I found this explanation extremely helpful when I did PCA in my postgrad multivariate analysis course: http://setosa.io/ev/principal-component-analysis/
191.
by alexcnwy 8y ago
Haha yeah I felt the same weird cult-like vibe. The thing about table-topics is you can't simulate it. People think they can do it until they get up in front of an audience of strangers. It's like that Mike Tyson quote: "ever
192.
by alexcnwy 8y ago
Is it possible that all the uber-style transport in a single city could be self-driving in the next 10 years? What would you say the probability of this happening would be?
193.
by alexcnwy 8y ago
Market share and data. I actually think the market share angle is far less important because the switching costs from the perspective of the rider are very low.
194.
by alexcnwy 8y ago
toastmasters - especially the improv "table topics" activities - has been indescribably valuable for me. It helped me go from being too shy to give a presentation in front of one person to giving talks to hundreds of strangers at
195.
by alexcnwy 8y ago
The third party doesn't know about the vulnerability. Company C posts bug bounty B in contract. Researcher X discovers vulnerability. Validator Y confirms the vulnerability and X gets paid (1-f)B where f is validator fee.
196.
by alexcnwy 8y ago
I've tried everything and I really like Dynalist (or any other outliner) as a personal knowledge repo / place to store notes and ideas. I find ideas do tend to be quite hierarchical and you can link between nodes if need be.
197.
by alexcnwy 8y ago
Very cool, great job! One bit of feedback: hover interaction doesn't make sense on mobile. Can tap to trigger hover as a workaround and use media queries to change instructions instead of telling user to hover.
198.
by alexcnwy 8y ago
The company posting the bounty. Third party verifies the bug. Why sarcastic?
199.
by alexcnwy 8y ago
I agree - it's a problem worth solving. Re: not having data to QA model, you can use the distribution of the generated labels to inform model updates. Alternatively you can get labels at the edge using a local edge interface then use t
200.
by alexcnwy 8y ago
Sounds like an actual use case for staking - bug bounties on teh blockchain!
201.
by alexcnwy 8y ago
Any suggestions on how to do that? Asking for a friend
202.
by alexcnwy 8y ago
I 100% believe that federated learning is going to be the new standard process in the future for many applications. Sending the model to the data instead of sending the data to the model (in the cloud) just makes so much more sense from a p
203.
by alexcnwy 8y ago
This approach fails e.g. The morphological operations you need for an infrared night time image with washed out white is different than what you need to detect the plate in the day time etc.
204.
by alexcnwy 8y ago
A related company that uses the blockchain in an interesting way for DRM (which I have my own views on) that was in the same Cape Town startup program as my startup: https://custostech.com/
205.
by alexcnwy 8y ago
This is troubling in the context of OpenAI deciding not to release their code and dataset for fear of it being put to bad use. It's a tricky topic but I get nervous at the idea of research being censored.
206.
by alexcnwy 8y ago
Be patient, it takes time for your body to recover from bouts of sustained exertion...
207.
by alexcnwy 8y ago
Paywall but can read article if you stop the page from loading before the modal pops up
208.
by alexcnwy 8y ago
If the words of some random person on the internet can dissuade you from being a founder then you're not cut out to be a founder.
209.
by alexcnwy 8y ago
Dynalist
210.
by alexcnwy 8y ago
Very cool and I think the product and landing page are gorgeous! One bit of feedback on the vid is I think it could be better timed to the music - the text zooms should be on the beat breaks.

More ›