goodmattg

5 ms·

1.
by goodmattg 3mo ago
Would it currently support Fable 5 via the restrictions Anthropic is placing on usage... because that's my major blocker
2.
by goodmattg 3mo ago
I flip back and forth between whoever currently has the more powerful frontier model that isn't cost prohibitive - subscriptions only, API pricing a non-starter. Today that's Fable 5 which has been excellent, as soon as it's
3.
by goodmattg 4mo ago
You can choose to be as charitable as you want with your lense on this. Incentives within all organizations lead to a certain class of worker taking over. From the "getting things done" perspective I like the lense of insurgents v
4.
by goodmattg 9mo ago
Fiction: A Little Life - challenging to read but exceptional Sci-fi Three Body Problem trilogy - forget the Netflix series the books are excellent, in my opinion #2 was the peak. The Glass Hotel, Station Eleven, Sea of Tranquility - anythin
5.
by goodmattg 1y ago
Perfect weekend read - kudos on a nice piece of writing
6.
by goodmattg 2y ago
I love this! looked into doing a similar project, you're competing against Hudl but using the phone instead of custom hardware (always preferred). Highlight segmentation may be a challenge with SOTA cv methods, but there are lot of dir
7.
by goodmattg 2y ago
sweet can't wait to read it! I'll also be at CVPR this year if you're presenting
8.
by goodmattg 2y ago
commercial license - the research group formed a corporate entity that licenses the body model and all derived work (SMPL-X, etc.): https://meshcapade.com/SMPL
9.
by goodmattg 2y ago
A few thoughts as I've done academic research & built products in this area: - if you're using SMPL body parameters this will have to stay research / open-source - is this leveraging some sort of monocular depth estimatio
10.
by goodmattg 2y ago
sell enough units to convince Khosla Ventures that you aren't just burning their money, forcing them to double down on the vision and bring in new investors for the next round. Rinse and repeat until the Large Action Model is either ex
11.
by goodmattg 2y ago
It's true, lack of highly available public transportation between the cities puts us behind other regions (not to mention other countries). I've been doing the ATX -> HTX drive for years. Best private alternative I've foun
12.
by goodmattg 3y ago
This is a stellar project
13.
by goodmattg 3y ago
Joining the nuance parade - Google expanded their entry points to Search years ago via Maps and the URL bar. Do I use Google traditional Google Search (google.com) to find things? Rarely. Do I use Google Maps to find bars / restaurants
14.
by goodmattg 3y ago
I also purchased and read this. If you enjoy the minutae of scientific advances you’ll love this, but not a page-turner. Also a helpful scientific / historical primer for anyone who wants to understand graphics (pixels, shaders, etc)
15.
by goodmattg 3y ago
This is great, but how is your performance on multi-table joins? We've been working on a homebrew solution using OpenAI internally conditioned on our schema, and can't bridge the multi-table gap. Also I'm skeptical if this ge
16.
by goodmattg 4y ago
I understand this is point-cloud diffusion but Ajay Jain et. al. (BAIR + Google Research) accomplished the first version I saw of this back in June with their Dream Fields paper (CPVR'22). As always this goes to show that if you can&#x
17.
by goodmattg 4y ago
Uses coordinate-based neural networks to model the scene volumetrically. However, in the case of this paper does not use an MLP to represent the scene. Instead, proposes to directly learn a voxel grid representation. For an excellent review
18.
by goodmattg 4y ago
my hypothesis: iced coffee
19.
by goodmattg 4y ago
Note of reference for readers who may not be as familiar. One (pytorch) process is: 1) Use torch.onnx to convert the model to ONNX 2) USe `onnxruntime` built for Web Assembly (WASM) to run the NN in the browser
20.
by goodmattg 4y ago
Need time to digest this paper, but you can assume if it's from Schmidhuber's group it will have some impact, even if only intellectual.
21.
by goodmattg 5y ago
in the vision group at ByteDance (TikTok)... and these kinds of mashups even blow me away. It's amazing to see what flexible tools, tight feedback loops, and global scale produce. reminds me of middle school where we would work togeth
22.
by goodmattg 6y ago
I was responsible for implementing Conjoint, Van Westendorp, and Gabor/Granger for a certain well known survey company. A few thoughts: Conjoint isn't typically used for pricing, but for finding the attributes of products that max
23.
by goodmattg 6y ago
As someone who came out of the same "prestigious and selective place as you", this take hits hard. I will say, part of what you learn doing research is if you even enjoy doing research. You play to win the game. Love the artwork!
24.
by goodmattg 6y ago
This is great!
25.
by goodmattg 6y ago
I can't tell if this is in jest... if you're actually interested, the role of HFT in markets is hotly debated. The HFT side argues that they improve markets by providing liquidity - and that's true to an extant, spreads have
26.
by goodmattg 6y ago
waxing poetic is one of my favorite hobbies. combination of climate, closer to hiking, and jobs
27.
by goodmattg 6y ago
COVID. they closed the Lake Austin location. There's still one on Soco
28.
by goodmattg 6y ago
My parents still remember when this happened! For me it was losing Magnolia Cafe at the Lake
29.
by goodmattg 6y ago
Native of Austin here, moving to SF. I can only speak to the trends I'm personally seeing, but the flood of tech money into Austin is sweeping my favorite parts of the city away. The places that filled the nooks and crannies in the cit
30.
by goodmattg 6y ago
This essay is both extremely poignant and nearly unintelligible, which is a vibe.

More ›