mcjkula Profile Banner
Maciej Kula Profile
Maciej Kula

@mcjkula

Followers
2K
Following
791
Media
34
Statuses
373

Keep learning about AI/ML and Physics. Happy to share my knowledge. ~ @learnbittensor

Virgo Supercluster, LG
Joined March 2024
Don't wanna be here? Send us removal request.
@mcjkula
Maciej Kula
2 days
The TAO halving happened at #7,103,976
6
15
136
@mcjkula
Maciej Kula
2 days
Happy Bittensor Halving everyone! 🥳
4
11
90
@learnbittensor
Learn Bittensor
1 month
🚨 TAO Emissions Got an Upgrade in Dynamic TAO How TAO gets distributed just changed. Instead of TAO emissions being determined by the EMA prices, the protocol now uses the net TAO inflows to decide which subnets gets "incentivized". The "old" way was to distribute TAO
15
42
161
@mcjkula
Maciej Kula
3 months
For anyone that wants to learn about DTAO (Injection, TAO/ALPHA halvings, ALPHA emission) and the "Subsidy Mechanism", that one is very likely for you. Thank you for having me @markjeffrey, was a great pod and hope it teaches some about how things work under the hood.
@markjeffrey
Mark Jeffrey
3 months
Hash Rate - Ep 131: Deep Dive on $TAO Emissions, Injections and the Halvening 🧙 Guest: @mcjkula There's a LOT of confusion about exactly how Bittensor injects liquidity, emits tokens, and even adjusts the rules based on certain conditions -- and what will happen at the TAO
13
19
105
@mcjkula
Maciej Kula
4 months
I've seen a lot of confusion about the subsidy mechanism, especially with the upcoming TAO halving, and thought it might be worth literally "dumping" some of my understanding and insights I have gathered recently. This will be a long one... 🥱 Where does the idea of the
30
26
140
@mcjkula
Maciej Kula
4 months
4
1
16
@learnbittensor
Learn Bittensor
4 months
🔄 Subnet Deregistration is Being Proposed Again for Dynamic TAO Subnet deregistrations are back on the table, and could free up TAO emissions and resources from inactive subnets with the introduction of BIT-0016. Here's how the proposal works: When Would It Trigger?
33
49
151
@JosephJacks_
JJ
5 months
It’s a privilege to learn. It’s an honor to teach.
0
3
68
@learnbittensor
Learn Bittensor
5 months
⛏️ TAOHash (SN14): Community-Owned Mining Pool Miners on subnet 14 have built a community-owned Bitcoin mining pool that peaked at 5 EH/s shortly after launch, representing 0.55% of Bitcoin's total network hashrate. Unlike traditional pools where operators keep all fees, ALPHA
4
18
109
@learnbittensor
Learn Bittensor
5 months
🖥️Chutes (SN64): Serverless AI Compute at Scale Miners on subnet 64 have built a serverless AI platform processing 160 billion tokens daily across 400,000+ users at 90% lower cost than other specialized platforms. Miners run "chutes" (AI models or computational tasks) and
0
19
121
@mcjkula
Maciej Kula
5 months
400K TAO seem to have appeared on root overnight, and it explains why the halving kept delaying as seen by some in the community. Here is the reason behind it: If you haven't been tracking it, the circulating supply had barely moved since over a month (only ~60K TAO increase,
Tweet card summary image
medium.com
On June 16th, I had a conversation with someone in the community who was confused by a shift in the halving date. This caught my attention…
14
24
182
@mcjkula
Maciej Kula
6 months
Day 1 of "Proof of Talk" was incredible! 🔥 Amazing conversations with many faces from the bittensor community, the energy here is truly unmatched. If we haven't met yet, shoot me a DM and let's connect tomorrow!
0
1
44
@mcjkula
Maciej Kula
7 months
I'll be attending the Proof of Talk event next week! Looking forward to connecting with everyone from the bittensor space - probably gonna feel like the imposter in the room though. It's kinda challenging for me, but I'm excited to meet everyone going and chat a bit. 👋
14
3
109
@mcjkula
Maciej Kula
7 months
You can now run @tplr_ai's Templar-1B model on @chutes_ai (SN64). That's cool! If you want to learn more about Templar/SN3 check out https://t.co/ovEba7UaNF 🫡
@learnbittensor
Learn Bittensor
7 months
📉 Templar (SN3): Incentivizing Loss Reduction The miners on subnet 3 just completed training a 1.2 billion parameter model through competitive loss reduction. Every 84 seconds, miners perform real training work on assigned data pages, measuring the model's mistakes (called
2
18
90
@learnbittensor
Learn Bittensor
7 months
📉 Templar (SN3): Incentivizing Loss Reduction The miners on subnet 3 just completed training a 1.2 billion parameter model through competitive loss reduction. Every 84 seconds, miners perform real training work on assigned data pages, measuring the model's mistakes (called
7
37
194
@JosephJacks_
JJ
7 months
Super excited to share that @mcjkula from @learnbittensor will be comprehensively covering the top subnets with detailed analysis -- publishing continuously. We are fortunate to have him on the team @latentholdings to accelerate the mainstream accessibility of Bittensor. 🧠
12
14
165
@learnbittensor
Learn Bittensor
7 months
✨ Learn Bittensor Got a Complete Makeover! Our platform just received a major redesign to make Bittensor more accessible than ever. 🎨 What's New: - Clean, modern interface with improved navigation - Enhanced search functionality across all resources - Better content
8
23
110
@mcjkula
Maciej Kula
7 months
I have to say, I am very disappointed by the recent quality of articles from @CrucibleLabs, specifically the misinformation in those, which I want to clarify and explain. The two articles, which catched my eyes were: 1. Liquid Alpha In this article it's being explained that
7
1
35
@JosephJacks_
JJ
8 months
Coming soon to a https://t.co/3486dUDpI6 terminal near you…
19
19
187