dheera

22 ms·

481.
by dheera 1y ago
We should support the rural lifestyle choice. For one, the food you eat comes from there.
482.
by dheera 1y ago
I haven't thought about an exact number, but I'd probably cave in for $20M. I'd finally be able to afford a house, never have to work for a toxic company again in my life, and could afford various preventative medical care wi
483.
by dheera 1y ago
> These people do not know the consequences of what they are doing. They might have known the consequences, but money is money. I feel like for 99% of people there is a certain sum of money for which they will give into pretty much any k
484.
by dheera 1y ago
I hope that this can one day be scaled, even if 100 years into the future. I do not want gold to be prized as a store of value. It is too useful as a material (inert, doesn't oxidize, food safe) that it would be vastly beneficial to so
485.
by dheera 1y ago
Chinese end user UIs tend to all be high-density per local user preference. Take a look at any of the mini-apps on WeChat or AliPay or pretty much any Chinese local app. Almost everything is linked within about 1-2 taps of the front page.
486.
by dheera 1y ago
The weird thing is, the biggest reason I don't use Cursor much is because they just distribute this AppImage, which doesn't install or add itself to the Ubuntu app menu, it just sits there and I have to do sudo chmod 755 pat
487.
by dheera 1y ago
Aw hell, more band-aids because people don't want to get software distribution done right. Can we please go back to the days of sudo dpkg -i foo.deb and then just /usr/bin/foo ?
488.
by dheera 1y ago
But will this run inside another docker container? I normally hate things shipped as containers because I often want to use it inside a docker container and docker-in-docker just seems like a messy waste of resources.
489.
by dheera 1y ago
The diagram is super vague. How are the lyrics encoded? What does the encoder look like inside? What is the input size, input format, output size, output format? Are the three encoder outputs added? Concatentated? When Mert and m-Hubert com
490.
by dheera 1y ago
I usually ask ChatGPT to "comment the shit out of this" for everything it writes. I find it vastly helps future LLM conversations pick up all of the context and why various pieces of code are there. If it is ingesting data, there
491.
by dheera 1y ago
I think this is going to be a problem as long as the metrics by which we evaluate students are intolerant of human errors and prioritize right answers over thinking skills. The current situation is that people need to pass exams, get certai
492.
by dheera 1y ago
Not really. I bounce from almost every website that gives me some GDPR, newsletter, or whatever popup. "Click to read full article?" F that, you should have pre-expanded it for me. I'm doing fine, I already have enough inform
493.
by dheera 1y ago
Academic "motivations" are just there to get papers accepted. Papers are just there to get tenure and graduate. Most of academia isn't after real science anymore.
494.
by dheera 1y ago
Venvs also suck because the user has to create and activate them. It would be nice if a foo.py file could just deal with all of that. A while ago I created this thing as a thought experiment. It likely doesn't work with more recent ver
495.
by dheera 1y ago
I rarely take pictures for memories, but rather to create art to share. If I think the photograph will be a work of art, then I'll go through the effort to capture it. If not, I just don't bother and focus on the experience.
496.
by dheera 1y ago
Got accosted by a popup when I entered the site so I bounced. Tip: If you want users to read your content, don't cover it up. https://i.imgur.com/uTmqaox.png
497.
by dheera 1y ago
To be fair, shared libraries have been problematic since the beginning of time. In the Python world, something wants numpy>=2.5.13, another wants numpy<=2.5.12, yet Python has still not come up with a way to just do "import numpy
498.
by dheera 1y ago
Seems everyone forgot HTML-SSI which worked something like this. Many servers and hosting websites of the 90s supported it.  Some content here <!--#include virtual=&quot
499.
by dheera 1y ago
Oh, sure, for the well-known models that are already on there. I just wish that new research would always spell it out in full instead of these silly block diagrams labelled with just e.g. "Cross Attention" and not the exact param
500.
by dheera 1y ago
Reminds me of this: https://www.youtube.com/watch?v=lGOofzZOyl8
501.
by dheera 1y ago
I think this is mostly the fault of RLHF over-indexing on pleasing the user rather than being right. You can system prompt them to mitigate this to some degree. Explicitly tell it that it is the coding expert and to push back if it thinks t
502.
by dheera 1y ago
Location: Currently bay area but looking to eventually relocate to one of { Seattle, Boston, LA } for lifestyle reasons Remote: Remote OK, Hybrid/In-Person OK if you have an office in one of the above cities Technologies: Gen AI, Multi
503.
by dheera 1y ago
Health insurance companies should pay for it. Their costs would come down if they subsidize the full R&D cost of this system for all car manufacturers. It would work in their favor. They're probably too stupid to think like that, t
504.
by dheera 1y ago
> employs a single-layer vector quantizer (VQ) codec and a single Transformer architecture to fully align I really wish when new models were released that they would draw a diagram of all the layers and the tensor input and output sizes
505.
by dheera 1y ago
Well sure but drones won't shit, that's why we need the organic piece. I guess they could drop rotten fruit in lieu of shit, but then we need a supply chain to restock the rotten fruit in the drones.
506.
by dheera 1y ago
Dooring is so incredibly preventable with simple computer vision and some kind of actuator that adds an audible alarm and mechanical 3x resistance to the door opening when a cyclist is detected. The door should still be openable in emergenc
507.
by dheera 1y ago
[flagged]
508.
by dheera 1y ago
I've actually been pondering using Gaussian splats for localization, I think it could be done. The idea would be looking for the pose that minimizes the MSE in density (rather than feature points or RGB similarity which are both vulner
509.
by dheera 1y ago
If the law set the speed limit as the hard boundary, with hard punishments for driving at 70, everyone else would drive at 65, occasionally getting close to 68-69. If the reasonable speed for a road is 70, the legal hard limit could be at
510.
by dheera 1y ago
I'm all for stronger enforcement of speed limits to save lives (e.g. ticket everyone over the speed limit with cameras automatically, no cops needed), but requiring shit to be installed into someone's car doesn't seem effec

More ›