| use std::str; | |
| fn main() { | |
| // -- FROM: vec of chars -- | |
| let src1: Vec<char> = vec!['j','{','"','i','m','m','y','"','}']; | |
| // to String | |
| let string1: String = src1.iter().collect::<String>(); | |
| // to str | |
| let str1: &str = &src1.iter().collect::<String>(); | |
| // to vec of byte |
| use borsh::{BorshDeserialize, BorshSerialize}; | |
| use solana_program::{ | |
| account_info::{next_account_info, AccountInfo}, | |
| entrypoint, | |
| entrypoint::ProgramResult, | |
| msg, | |
| program_error::ProgramError, | |
| pubkey::Pubkey, | |
| }; |
Picking the right architecture = Picking the right battles + Managing trade-offs
Artificial Intelligence (AI) has evolved rapidly, with language models like GPT-4 capturing attention. However, the real future of AI lies in embodied intelligence—systems that can interact with the physical world through robotics and sensory perception. Unlike disembodied language models that operate in digital spaces, embodied AI must navigate complex environments, interpret sensory data, and perform physical tasks. This shift towards embodied intelligence opens the door to groundbreaking applications and significant economic impact.
Developing embodied AI systems is far more complex than working with traditional language models. Embodied agents need to:
| <!DOCTYPE html> | |
| <html lang="en"> | |
| <head> | |
| <meta charset="UTF-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> | |
| <title>Game</title> | |
| <style> | |
| body { | |
| margin: 0; | |
| overflow: hidden; |
| import numpy as np | |
| import pandas as pd | |
| import matplotlib.pyplot as plt | |
| import scipy.optimize as sco | |
| from quotes import tickers | |
| df = pd.read_pickle("quotes.pkl") | |
| logrets = np.log(df / df.shift(1)) | |
| rmean = logrets.mean() * 365 |
