Tweet fixado
jbreak
1.3K posts

jbreak
@jbreak_codes
Dev. Research/Edu Threador, love details. @sapiencemarkets | @Nethermind. Master's at @UofT.
depths of rust codebase Entrou em Şubat 2024
195 Seguindo468 Seguidores

calling Coinbase engineers
when I was reading through the whitepaper I was constantly thinking about the fact that credit cards exist in order to protect the customers
now add security lacking agents on top of them being able to make a no-refund payment
x402 has long ways to go
Serpin Taxt@serpinxbt
no one has explained how x402 is going to handle fraud, OFAC sanctions, AML, and typical KYC constrains of running a business that accepts payments agentic payments will be the #1 psyops for fraud “but it’s just an agent bro!! he’s just buying APIs” - some guy in north korea
English

@ngocquang19877 @WazzCrypto @PatrickAlphaC @AnthropicAI @OpenAI Turtles all the way down
Bette protects against accidental leakage from LLMs that just read .envs
English

@jbreak_codes @WazzCrypto @PatrickAlphaC @AnthropicAI @OpenAI this is not better than a .env on your own machine. You need to communicate with the Vaults using encryption and guess what ? the pem files need to store on your file.
If you do things on your computer, using Apple ID and Apple vaults are way way better
English

DEAR @AnthropicAI @OpenAI
CAN YOU PLEASE HARD-CODE YOUR MODELS TO STOP RECOMMENDING PEOPLE STORE THEIR KEYS IN PLAINTEXT

English

@WazzCrypto @PatrickAlphaC @AnthropicAI @OpenAI I believe so, yes. When quickly iterating and having to push code here and now you can do .env, but anything that goes into prod with malicious eyes potentially trying to break it should be using this IMO
English

@jbreak_codes @PatrickAlphaC @AnthropicAI @OpenAI i guess this would also be the best solution for API keys?
English

@WazzCrypto @PatrickAlphaC @AnthropicAI @OpenAI HashiCorp Vaults or cloud-native secret management solutions
English

@Evan_ss6 citrini thesis is based on AI capabilities accelerating infinitely with more spending. if the AI acceleration sigmoid is real (99% sure it is since it doesn't have a working real-world model although it does seem like it does), the thesis breaks down and we are going to be OK
English

@neetcode1
Hey sir there is a bug in the solutions for this problem:
neetcode.io/problems/searc…
3rd solution uses an optimization that counts refs for each node. When removing a word, the refs are only updated in the leaf node and not along the path, which does not improve runtime
English

Hey @Polymarket I'm sorry but what the actual hell
Do you actually pay people to do this for a living?

English

@Arvolear the one that you write yourself
(shameless plug incoming)
dl.acm.org/doi/pdf/10.114…
English

@123skely @jbreak_codes wish you didn’t end with the correct explanation. would’ve been peak lil shid tag material
English

@cryptobuilder_ There is no industry, I love you
(gambling if we are talking web3)
English

@jbreak_codes What is ‘our industry’ these days?
English








