Vivek Tanna

@pashov Inshort you are also not sure.

Everyone is asking "what to do in web3 security in the AI era" I've been observing multiple masters of web3 security and their AI usage. While everyone starts with just "summarise/explain the codebase", mostly everyone ends up building their own toolings. Tools are usually vibecoded Python & Bash scripts plus Markdown files for the AI. It's packaged expertise. Why would you build such tools? Because they do in 1hr what you did before in 5 days or more. Still, many of the great auditors are not all-in into AI. Many of the top 1% are using A LOT of AI, but their own judgement is still driving the car. Whatever works - that's the right solution. You need to have the correct metrics, KPIs, Key Results or whatever you want to call them. Measure rigorously, constantly and iterate. You must find more and better findings than others, faster. Do what works. Now go🫡

@kalashvasaniya @elonmusk maybe but codex 5.5 >> composer 2.5

i dont believe im gonna say this but composer 2.5 >>>>>>>>>> opus 4.7 @elonmusk cooked

@MonkeyMeaning Seriously good to read, but as suggestion cant u put this as X blog much more easier to read. Long thread are boring tbh.

And this is how AMMs work.

Say, we live in a world where we buy oranges in exchange for apples, and vice-versa. A vendor has 100 oranges and 20 apples in his basket. We're buying oranges from him and we're paying him in apples. How many oranges is one apple worth?

@MatiVivorita Software still needs thinking. Agents automate the manual work, but not the taste, the design, the saying no.

@7a7zz The good thing is both where short so fastly get all the conclusion. But from which tool u generated that blog imgs

published 2 back to back banger articles. there is absolutely no way Im topping this oml

@SahilPanhotra @forgebitz Thx to @thebuggeddev youtu.be/uqpXAfNEY4g?si…

@SahilPanhotra @forgebitz How ?? Any skills? Prompt? Or anything you like to put lights on ?

i don't know a single developer using gemini models

would be over $300M if include @DriftProtocol 👀

LOL codex always gives some new and unique names to its subagents when it spawns

@nexxeln @opencode That's the best ease of doing one can do.

if you’re in India, subscribing to @opencode just got a lot easier upi autopay is now live for opencode go ₹900/month for generous limits and reliable access to open source coding models

@MonkeyMeaning @AmpCode @claudeai AMP inbuilt comes with various tools pre set up, even though underneath it uses opus 4.5 and codex mostly. but ya im too amazed with the output

switching from @AmpCode to the $200/mo @claudeai plan for a month has taught me an important lesson just use the $200/mo to purchase amp credits

@lucidzk Bro ur feed 🤣🤣

zcash unstoppable private money.

@christianascone Thx ! Will for sure check it

@WizzOnChain_ There's an experimental version for Windows in progress: github.com/tw93/Mole/tree…

I'm trying Mole for the first time and wow! It just found 10GB of trash to clean 🗑️

@beeman_nl Yup yup i tried for me 5.2 xhigh is like mature (but slow af) not 5.3 fan. 5.4 found bit more balanced but still gives psycho endings.... (Not sure about this fillers)

@WizzOnChain_ Really? Did you see GPT5.4 in Codex use conversational fillers? I haven't found any.

Previous models from OpenAI would utter sycophantic stuff like "I'm so glad I was able to help, let me know if I can blablabla....". Not GPT-5.4: 🔥

@mattjay 😂😂

TL;DR - guy thinks he hacked Perplexity. Perplexity responds with "lol, no our billing is just async. check your inbox." Guy finds his bill: *shocked pikachu*

@0xMahalakshmi True for most of the web3 events nowadays

ETH Mumbai in a nutshell

@PatrickAlphaC @mroliverpt looks kinda cute bro 🤣

@mroliverpt Bro lol

how do i explain this to friends when they ask what i do

@Bhavani_00007 It simply filters at the client side, so technically legal(or grey area)

Brave blocks YouTube ads, but how does it do that legally? 🤔