chaos_emergent

9 ms·

91.
by chaos_emergent 2y ago
what about a mechanism that 'locks' the trackball in place at rest, and then requires a little bit of force to 'unlock' the trackball to track?
92.
by chaos_emergent 2y ago
I have a moonlander that I love, and I want to add a trackball or trackpad to it. I saw the mod that Kevin Christensen wrote on ZSA's site: https://blog.zsa.io/diy-moonlander-trackball/ The missing piece is that t
93.
by chaos_emergent 2y ago
We're neither disembodied brains nor bodies in isolation of basic needs like shelter or desires for energy-intensive conveniences, so I challenge that we should try to assess energy consumption in the way you're proposing. The out
94.
by chaos_emergent 2y ago
You can try it for yourself, my company is in my profile.
95.
by chaos_emergent 2y ago
You should make it something random like “pineapple”
96.
by chaos_emergent 2y ago
I agree! I think it’s “worth it”. A parallel: it costs an OOM more to drive places than, say, bike places, but it allows us to do the things we did before an OOM faster and expands the types of things we can do. You can make the argument th
97.
by chaos_emergent 2y ago
I’d have to give away some of my secret system prompt sauce to tell you, but I’m sure that the summarizations I get are solid, we’ve done extensive evaluation of truthiness. I have tens of thousands of hours of calls summarized monthly from
98.
by chaos_emergent 2y ago
By that argument, anything that an AI system could do today that a human does right now is an incredibly massive waste of our resources :-)
99.
by chaos_emergent 2y ago
It sure does beat needing to have someone to do it for you :-)
100.
by chaos_emergent 2y ago
Yeah, an incredibly dumb analysis. I did a back of the hand estimate based off of the following assumptions: gpt-4o: 10 H100s = ~7000W power requirement, 30 seconds to generate a summary for an hour-long call human: ~10.7 kWh/year ener
101.
by chaos_emergent 2y ago
I’m confused by the evidence they use: > to justify those costs, the technology must be able to solve complex problems, which it isn't designed to do Planning and reasoning are the two greatest areas of research in AI right now, wit
102.
by chaos_emergent 2y ago
specifically, it's 20x a GHG compared to C02, and I don't really know what you mean by it being part of carbon cycle.
103.
by chaos_emergent 2y ago
What do you think that is?
104.
by chaos_emergent 2y ago
actually I just went into the paper and skimmed it. Their image tokenizer creates image encodings in a discrete space of dimension 8192; each image consists of 1024 discrete tokens. The 8192 vector size matches their language tokenizer (dis
105.
by chaos_emergent 2y ago
can someone give an intuition for the unified tokenizer that they mentioned? I understand the idea of using a GAN or an autoencoder to make a image tokenizer, but it's unclear how you combine that with text to have a single tokenizer o
106.
by chaos_emergent 2y ago
They're releasing a model to the public that likely costed tens of millions, for free, and you're complaining that they don't want to have to deal with bullshit PR for giving a teenager fucking around in his backyard the reci
107.
by chaos_emergent 2y ago
orders of magnitude
108.
by chaos_emergent 2y ago
When training a large model you're doing a single forward pass + backprop over multiple Infiniband-connected nodes for a single model instance , so if one node goes down it takes a logical unit of nodes down with it. For reference, GP
109.
by chaos_emergent 2y ago
If you don’t think an AI could do that then I think you need more imagination and to follow the OOMs
110.
by chaos_emergent 2y ago
This is the most insightful nugget into OAI's consumer product strategy I've read, thank you.
111.
by chaos_emergent 2y ago
have you tried using gpt-4o? It's pretty incredible at recognizing handwriting.
112.
by chaos_emergent 2y ago
Maybe instead of overloading the shorthand with two definitions, it's best to just use the actual words.
113.
by chaos_emergent 2y ago
as a follow up curiosity, has anyone tried using LoRA on the entire model for pretraining to compare regular training model performance to LoRA?
114.
by chaos_emergent 2y ago
The findings are that the best fine-tune performance comes from fine-tuning all weights, followed my MLPs, followed by attention heads, using LoRA. Authors assert that the performance difference is based on the target module of the NN. Isn’
115.
by chaos_emergent 2y ago
Isn’t this fairly obvious after a two second glance at the abstract
116.
by chaos_emergent 2y ago
> Every time i've actually tried something like this it has not outperformed statistical process control. There are clear counterexamples to your experience, most notably in maintaining plasma stability in tokamak reactors: https:&
117.
by chaos_emergent 2y ago
"Why would you even try to predict the weather if you know it's going to be wrong?" - most OCs on this thread
118.
by chaos_emergent 2y ago
For a website named “reason.com” this article lacked it. A link to a poll and a WSJ article, and otherwise drivel lamenting the change in habits and attitudes toward play. There is a very good case to be made against the impacts of social m
119.
by chaos_emergent 2y ago
Drawing bounding boxes around license plates, and then running OCR on those license plates, is a relatively solved problem. My guess is that low confidence predictions will probably still be manually reviewed.
120.
by chaos_emergent 2y ago
Let’s say that the city of LA theorizes that enforcement of parking via these cameras will be cash flow positive; assuming the $11M contract size, 100 cameras and 5 years at face value you’d expect each device to bring in $60/day to br

More ›