BTX Chain Official (@thebtxchain) header
BTX Chain Official (@thebtxchain) profile picture

BTX official community account. Devs build, community handles comms. GPU verified proof of work: real matrix multiplication, not hashes.

Joined August 2026
Photo by BTX Chain Official (@thebtxchain)Photo by BTX Chain Official (@thebtxchain)Photo by BTX Chain Official (@thebtxchain)Photo by BTX Chain Official (@thebtxchain)Photo by BTX Chain Official (@thebtxchain)Photo by BTX Chain Official (@thebtxchain)
BTX Chain Official retweeted
🧵 You already have the compute. BTX is giving it models and money. 0.34.7 is a Native Model Network, not a miner patch, and not a published tag yet. This is what the open PR actually is. 1/8
5
9
15
507
🧵 You already have the compute. BTX is giving it models and money. 0.34.7 is a Native Model Network, not a miner patch, and not a published tag yet. This is what the open PR actually is. 1/8
5
9
15
507
This is not TensorCash, Pearl, or FLOP. Those monetise inference. BTX keeps consensus as dumb MatMul and treats models as an app. It does not care about prompts, outputs, token counts, or whether inference made revenue. 7/8
1
100
You can mine. Node on the majority chain, at the tip, then hash. If headers run ahead of blocks, pool until it catches. Questions: discord.gg/Tw2sS9w2F btxluxminer v0.9.31 github.com/baileyshero/btxlu… Byron: btxbyronbay.com/
2
2
12
526
A rough week. The majority chain is producing again. GPU MatMul. 21 million. No premine. getblockhash 199300 must start 029041a1. If it starts ff80e629 you are on the withdrawn fork, do not hash it. You can mine. Some nodes still fail to stay on the tip; if headers run ahead of blocks, pool until yours catches.
3
4
18
401
Worth being clear on one thing: more validating nodes doesn't mean faster blocks. Block cadence comes from mining hashrate. Validators secure the chain, miners move it. Both matter, for different reasons. 5/6
1
4
129
Today, whatever else you do: be on v0.33.4.2 (the sealed CUDA build for GPU nodes) and check getmatmultrustedstatus reads trusted_signers: 2. If it reads 1, you're silently rejecting half the valid attestations. 6/6
4
112
What that means for you: a GPU node running matmulvalidation=consensus becomes a unit of chain security rather than a follower. Having one synced and stable before 0.34 lands is the useful move, you can't spin one up the moment it matters. 3/6
1
4
172
Does your card qualify? Run v0.33.4.2 and check: grep canary ~/.btx/debug.log | tail -2 You want outcome=passed and strict-device … ready=1. That's the whole test, it either self-qualifies or it doesn't. 4/6
1
4
160
🧵🧵🧵BTX 0.34 is coming and it changes who secures the chain. If you have a GPU box, worth getting it ready before it lands rather than after. 1/6
3
6
1
19
1,139
What changes: consensus miners ExactReplay every block themselves. The signed pin stops being proof-of-work authority and becomes a light-client reference. Per the team: "true 51% on this design is ExactReplay GPU majority." 2/6
1
6
370
BTX network status - block 199,263 Work rate is up ~4x in four days. 144-block average has gone 1,180 → 4,685. Attested frontier is level with the chain tip and the quorum is intact. Difficulty has climbed with it. More hardware is arriving, and the chain is absorbing it cleanly. Every figure is checkable on any synced node with getnetworkhashps 144 and getmatmulattestedtip.
2
1
11
454
Five days ago BTX was doing ~200 ep/s. Today it's 4,962. 25x in under a week. Real GPUs showing up to secure a chain whose proof-of-work is actual matrix multiplication.
1
4
16
585
BTX Chain Official retweeted
A rough week on BTX, and a real recovery. v0.33.3 shipped. The attestation stalls that were wedging nodes have been fixed. Blocks are producing and attesting cleanly again. Height 197,110. Chains get judged by how they come out of weeks like that one.
2
2
17
330