Ben

breaking things 24/7

[Ben @ffsxdev]

@showxlate @in19h on a real on a realest homie mcbrodie you drop in a binary and it tells you what architecture it is.
there's 1 button and 4 words on the screen - it doesn't need revolutionary intentional design, it's a side project.

[Ben @ffsxdev]

@in19h Revolutionary tooling, though I believe I'm now out of a job

[Ben @ffsxdev]

@BrudaSol @K3shavGupta It's running using Mac Catalyst, with a Frida hook so the anti-analysis framework will run the binary even though the code has been modified

[Ben @ffsxdev]

I played Jev in Clash Royale, and it nearly beat me.

I've seen a lot of demos recently where people have Jev play games, but most of them were singleplayer, and some paused the game so Jev had time to react.

I extracted the game state from the live game - and mapped it into questions that Jev could answer, then wrote an interaction layer to perform the decisions that it made.

The result, a freakishly good, real-time Clash Royale bot. Huge props to the team @typesafeai on the release.

[Ben @ffsxdev]

@apaz_cli Maybe.. I'm Arena 21 on my main 🤷‍♂️
Still crazy that an AI can do this in real time.

[Ben @ffsxdev]

I extracted the game state with a Frida hook that found the current game state and used offsets I found in IDA to read all of the current values, polling at 10hz.
Interaction I did by injecting UIKit taps into the app, so it didn't have input delay & broken taps from following my cursor

[Ben @ffsxdev]

@slendidev PoW makes it more expensive at scale, not for one person running one headless browser. It's aimed at people firing hundreds of thousands to millions of requests at a website, and it's rather effective at deflecting that.

[Ben @ffsxdev]

@user9JKsuG Those algorithms have been open source for that version for over a year, and you still managed to get them wrong - and not include Argus.
Congratulations on your slop.

[Ben @ffsxdev]

@boyuan_chen @oskargroth As soon as it hits fairplay obfuscation it'll crumble. Can't deal with the massive dispatchers

[Ben @ffsxdev]

@precious_tagy @iamdanztee Build it right the first time, so you don’t have to build it again.
Not building with the future in mind will just create more and more technical debt

[Ben @ffsxdev]

@precious_tagy Instantaneous streaming to 3 devices is cool, but not difficult. Scaling that to support hundreds of thousands, if not millions of users globally is the real challenge

[Ben @ffsxdev]

Session hijacking is where you authenticate with a session token that isn’t your own, and has been stolen from a user via a MiTM attack, or even sometimes through prediction (if you have a JWT secret key, you can authenticate as any user)

What you have done here is copy and paste your cookies from your browser into your IDE, creating a glorified API client.
The API you gain access to is also not the V2 paid api, as it’s missing certain filter terms like “has: cashtag”. It’s also missing the websocket stream, which imo is the most valuable part of their paid API.

[Ben @ffsxdev]

@lusongg @Armv7lFx looks like Vultr by the styling

[Ben @ffsxdev]

@Daybreak0x @ProphetBots. Nothing else can compete

[Ben @ffsxdev]

@JuicyCandle @BrwnPB @ProphetBots Sorry, C# isn't used in our stack.

[Ben @ffsxdev]

Transaction execution is integral to any crypto project, and when you are working with something as complex as AutoSniper, it has to be perfect.

🧵My Journey Perfecting Transaction Execution for @ProphetBots

During AutoSniper development, we realised that bundling was vital to ensuring the best entries and reducing RnG.
With 200+ entries being performed all at once, entries would be at the mercy of validators, with people being left in the dust because of a 300-millisecond delay.

ExecutorV2 was born and has seen some extremely good entries during testing
However, this meant that manual transactions did not receive the benefits of bundling.
We didn’t want manual buys / sells to not have the benefit of bundling where Prophet wins every block, leaving standard transactions pushed to the chain and sometimes waiting for multiple blocks to pass before a confirmation is seen.

This is why Project Horizon was created.

Lightning-fast confirmations. Block 0 entries. Higher Block Positions than all the competitors.
For the entire Prophet Ecosystem.

———— Reaching Block 0 ————

Block 0 is the block immediately after a signal is created. We can’t use that as a target though, as the signal algorithms need time to analyse a token, holders and volume, and check it across hundreds of different criteria between it is posted to holders.

How do we quantify speed then? We have to alter the definition of ‘Block 0’ slightly to match what we’re looking for with our entries, which is the block to which our Signal is posted for users.

The goal was to make sure that by the time you are notified of the signal, your entry is already waiting for confirmation on the chain.

This isn’t a straightforward task, but reaching it would set us apart from every competitor out there.

Collectively, @antarctic, @permarhino and I worked together to cut down the latency between the validation of a signal and the request for transaction creation to under 300ms.

We transitioned all our systems to use a queue that is hosted on our local node, to reduce the communication time between different Prophet systems to under 7ms. @antarctic and @permarhino began to introduce parallel on-chain simulations to fetch on-chain data and started to pre-fetch and cache data to reduce both our overheads and our processing time in the auto sniper.

Then, all the pressure was on Horizon
- Integration of queue systems to reduce confirmation time
- Caching important chain values like gas price
- Adding a prediction algorithm to reduce queries to chain

I was able to reduce the time for a transaction to be broadcast to builders to under 300 ms.
This meant that our entries' 99th percentile was Block 0 after we posted the signal.
When a signal is posted, there's no time to waste and every millisecond counts. This reduction allows us to have the fastest time-to-entry in the game.

——- Optimising Block Positions ——-
Block Positions are one of the most important things in crypto trading, as they determine your entry price.
When you have over 200 people entering into one signal, you can often get dragged behind and get a worse entry.

Normally, when implementing bribes, they only benefit the position of the entire group, meaning if you bribe 0.04 ETH, that benefits everyone else in that bundle as it doesn’t affect your position inside the bundle - just the bundle’s position in the block.

That’s why we knew we needed a priority system within Horizon.

Horizon employs a system that guarantees that the highest payer will get the best entry. There’s no more RnG in trading, it’s entirely PvP.
This PvP is only within the Prophet AutoSniper, and even without a bribe, you’ll still beat out all the competitor’s entries 😉

It’s coming, and it’s going to be better than ever.
Don’t fade - @ProphetBots 🚀

[Ben @ffsxdev]

@BrwnPB @IYKYK170 @ProphetBots @antarctic We could make that happen 👀

[Ben @ffsxdev]

@growing_daniel localhost:5000

[Ben @ffsxdev]

@TenZOfficial no