[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![KaspaFacts Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1749792064827457536.png) KaspaFacts 𐤊 [@KaspaFacts](/creator/twitter/KaspaFacts) on x 5113 followers
Created: 2025-07-23 12:19:30 UTC

#Kaspa and AI

💡TLDR:  Don't use AI to do your thinking for you.

I keep seeing more and more AI generated posts and responses about Kaspa. (It appears Bitcoin Maxi's are using AI to do all their thinking for them, they post long responses with very little meaning)

If you are using AI for your information, and you cannot recognize when it is wrong, you are going to be working on bad information.

🟥EXAMPLE:  Yesterday I was using an LLM to dig more into Kaspa.

When on the subject of tiebreaking blocks with the same blue work, it insisted that comparing hashes in lexicographic order can be different from numerical value (ties are broken by lowest hash),

✅Process:

1⃣Byte-by-byte comparison:  Starting from index 0, each byte is compared using standard integer comparison.

2⃣First difference wins:  As soon as two bytes differ, the hash with the smaller byte value is considered "less than" the other.

3⃣Complete equality check:  Only if all XX bytes are identical are the hashes considered equal.

It then proceeded to give me several incorrect examples, I had to correct the example, the LLM agreed its math was incorrect, then provided a new incorrect example.

After being unable to provide a working example or a full explanation of something as simple as tie breaking blocks by hash, I gave up.

This morning I revisited this problem, I got it to retrieve the relevant information from the codebase (I still can hardly read Rust btw, so I have to assume the LLM got this part right), then walked the LLM through the process, step by step, then finally got the answer I was looking for.

✅The missing component:

Hash is stored in Little-endian format.

✅Result:

Hashes in Little-endian compared from index X lexicographically can differ from the numerical comparison.

Kaspa breaks ties in a way that differs from "lowest hash", but still provides deterministic results as if "lowest hash" was used.

(I still have to assume this information is accurate since I cannot read Rust, and I also assume it is simply a hardware/software optimization)

![](https://pbs.twimg.com/tweet_video_thumb/GwirHC2agAEUnjj.jpg)

XXXXX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1947994982981288417/c:line.svg)

**Related Topics**
[generated](/topic/generated)
[coins ai](/topic/coins-ai)
[kaspa](/topic/kaspa)
[coins layer 1](/topic/coins-layer-1)
[coins pow](/topic/coins-pow)
[coins made in usa](/topic/coins-made-in-usa)
[bitcoin](/topic/bitcoin)
[coins bitcoin ecosystem](/topic/coins-bitcoin-ecosystem)

[Post Link](https://x.com/KaspaFacts/status/1947994982981288417)