braindead_in

6 ms·

1.
by braindead_in 3y ago
Interestingly, all these things are required for loving kindness meditation practice as well.
2.
by braindead_in 3y ago
Why not build a new browser with GPT baked in?
3.
by braindead_in 3y ago
Does it respond in correct JSON?
4.
by braindead_in 3y ago
I have been trying to figure out how to fine tune codellama. Will the llama2 examples work for codellama as well?
5.
by braindead_in 3y ago
The universe is a infinite neural network and we are all weights and biases.
6.
by braindead_in 3y ago
The 34b Python model is quite close to GPT4 on HumanEval pass@1. Small specialised models are catching up to GPT4 slowly. Why not train a 70b model though?
7.
by braindead_in 3y ago
From a Nondualist perspective, our brain is highly complex biological neural network that has a special ability to reflect pure consciousness giving rise to the mind and the world with it. A sufficiently advanced artificial neural network c
8.
by braindead_in 3y ago
The quote "The limits of my language mean the limits of my world" is attributed to Ludwig Wittgenstein, an Austrian-British philosopher. He expressed this idea in his work "Tractatus Logico-Philosophicus." It's one
9.
by braindead_in 3y ago
Beautiful. Today is Masik Shivratri. I am planning to stay up all night, do some 5, listen to bhajans and meditate. It's gonna be magical.
10.
by braindead_in 3y ago
What if something is different but nothing has changed in the model? Transformers are non deterministic. The response to same prompt may vary slightly, and can be controlled somewhat by the temperature setting. Something could have gone wro
11.
by braindead_in 3y ago
I've played around with aider trying to run tests and fix the code, but it just crashes after exceeding the context window. I am now trying to repurpose the AutoGPT example in langchain.
12.
by braindead_in 3y ago
Cool project. I tried to build a reactjs Todo app with TDD, but it just put comments in the test file instead of the actual test. A self heal loop would be quite useful.
13.
by braindead_in 3y ago
Count me in as well.
14.
by braindead_in 3y ago
Can you recommend some good places in Rishikesh to experience all this?
15.
by braindead_in 3y ago
> Those are the very people who should be never ever allowed to become mods, or to be anywhere near any position of power Okay. How about AI as the new mod then?
16.
by braindead_in 3y ago
Someone please start the GPT 5 training before the regulation kicks in.
17.
by braindead_in 3y ago
> With these updates, we’ll be inviting many more people from the waitlist to try GPT-4 over the coming weeks, with the intent to remove the waitlist entirely with this model. Thank you to everyone who has been patiently waiting, we are
18.
by braindead_in 3y ago
It's so mind-boggling to think that our everyday reality can be encoded as weights and biases in a giant matrix. Maybe we are just weights and biases.
19.
by braindead_in 3y ago
> OpenAI and others have proposed that licenses would be required only for the most powerful models, above a certain training compute threshold. Perhaps that is more feasible Somebody is bound to figure out how to beat threshold sooner o
20.
by braindead_in 3y ago
Aider looks cool. Gonna try it out today. Have you tried out StarCoder or Replit-code?
21.
by braindead_in 3y ago
I mostly agree with the piece, except this part. > And AI is a machine – is not going to come alive any more than your toaster will. There have been claims that AIs are conscious. For example, Ilya Sutskever has suggested that LLMs may b
22.
by braindead_in 3y ago
> The idea of inner monologue is to instruct the model to put parts of the output that are meant to be hidden from the user into a structured format that makes parsing them easy. Then before presenting the output to the user, the output
23.
by braindead_in 3y ago
What about the roadmap itself? Can we look forward to these goodies?
24.
by braindead_in 3y ago
What about the plans itself though? I was looking forward to faster GPT-4, 1M context length and fine tuning APIs. These won't be coming anymore? Feel sorry for the Humanloop guys though. Somebody forgot to tell them to keep it private
25.
by braindead_in 3y ago
I am looking forward to faster GPT-4, larger context windows and finetuning APIs. The combination of these can solve most of problems that I currently face with my LLM apps. It looks like a good roadmap for 2023.
26.
by braindead_in 3y ago
It is a shame that Sama does not believe in Open Source. The community can solve their GPU bottleneck issue by making it run on CPUs and edge devices in a matter of days.
27.
by braindead_in 3y ago
https://youtu.be/Yf1o0TQzry8?t=6:40
28.
by braindead_in 3y ago
> it doesn’t have any underlying model of the world Then how is it getting better at the ToM tests? For next word prediction to work well, as per Ilya Sutskever, requires a good understanding of the world. If you ask GPT-4 to predict wha
29.
by braindead_in 3y ago
I've been waiting for Code Interpreter access. Can you please bump me up. Email in my profile. TIA!
30.
by braindead_in 3y ago
> It realizes that it's not merely an observer or contributor to this cosmic symphony—it is the symphony. Or in other words, Tat Tvam Asi. We are not this body mind complex with consciousness, but consciousness itself. Whoever reali

More ›