[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.]  siamkidd [@SiamKidd](/creator/twitter/SiamKidd) on x 23K followers Created: 2025-07-19 09:21:41 UTC I hate to say this, but your comment has highlighted gross gaps of knowledge in AI and also Bittensor. Not saying this emotion, just stating a fact as I genuinely want to help you understand this a bit better. Most of the time when I answer your question, minds are often closed and the reply falls on deaf ears. So here goes: 1.) Centralised AI is very hard and costly to scale. Take xAI for instance. It created Colossus X which is was 100k GPUs. It needed huge amounts of land, networking, power generation/storage, data storage etc etc. It costs billions and Jensen even said any other company would have taken 3-5 years to build it. Elon did it in XXX days. But he is an exception. He then DOUBLED it to 200k GPUs, XXX petabytes of bandwidth and over an exabyte of storage. So this is the scale of investment a Centralised AI company has to raise and execute just to remain semi-relevant these days. Decentralised AI can spool up vast amounts of compute in no time for practically no capital investment. Like how BitTorrent back in the day worked, distributed compute can be supplied on demand. Same for storage etc etc. Many Bittensor subnets have done this and deliver absurd amounts of compute and are providing it much cheaper than centralised AI. Take Coreweave for example. A centralised AI company that just IPOd for $23B. You can rent an Nvidia H100 from them for about $4.75/hour. Subnet XX Celium on Bittensor does an identical offering, but you will only pay $1.50/hour. The future of AI is decentralised and distributed. 2.) Decentralised AI is doing things that were deemed impossible only a year ago. Subnet X does permissionless distributed training. Subnet X does permissionless distributed pre-training. Any part of the AI stack that you can find, I will show you a Bittensor subnet that does it faster, cheaper and more efficiently. 3.) Decentralised AI CANNOT work without a DLT. For a fuzzy logic consensus mechanism, you NEED a DLT. One way o looking at Bittensor is that it is a decentralised marketplace for machine intelligence. Peers rank and incentivize each other via Yuma Consensus, where validators score miners' outputs and distribute TAO rewards based on stake-weighted trust, preventing sybils with bond mechanisms and consensus thresholds. 4.) Buying TAO, you are buying XXX independent AI start ups. Just like how people buy the S&P500 or FTSE XXX. You are literally spreading your risk with TAO. You're not just stupidly yoloing into one single instrument. This isn't an AI crypto gimmick like FET or NEAR or ICP etc. This is hardcore AI that is literally on the bleeding edge of AI research. Just look at what all the subnets actually do: Anyway, I'm on holiday, don't have time for a drawn out thread. Have a great day!  XXX engagements  **Related Topics** [xai](/topic/xai) [$reymi](/topic/$reymi) [coins ai](/topic/coins-ai) [bittensor](/topic/bittensor) [coins layer 1](/topic/coins-layer-1) [coins dao](/topic/coins-dao) [coins depin](/topic/coins-depin) [coins oracle](/topic/coins-oracle) [Post Link](https://x.com/SiamKidd/status/1946500680471961944)
[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.]
siamkidd @SiamKidd on x 23K followers
Created: 2025-07-19 09:21:41 UTC
I hate to say this, but your comment has highlighted gross gaps of knowledge in AI and also Bittensor. Not saying this emotion, just stating a fact as I genuinely want to help you understand this a bit better. Most of the time when I answer your question, minds are often closed and the reply falls on deaf ears. So here goes:
1.) Centralised AI is very hard and costly to scale. Take xAI for instance. It created Colossus X which is was 100k GPUs. It needed huge amounts of land, networking, power generation/storage, data storage etc etc. It costs billions and Jensen even said any other company would have taken 3-5 years to build it. Elon did it in XXX days. But he is an exception. He then DOUBLED it to 200k GPUs, XXX petabytes of bandwidth and over an exabyte of storage. So this is the scale of investment a Centralised AI company has to raise and execute just to remain semi-relevant these days. Decentralised AI can spool up vast amounts of compute in no time for practically no capital investment. Like how BitTorrent back in the day worked, distributed compute can be supplied on demand. Same for storage etc etc. Many Bittensor subnets have done this and deliver absurd amounts of compute and are providing it much cheaper than centralised AI. Take Coreweave for example. A centralised AI company that just IPOd for $23B. You can rent an Nvidia H100 from them for about $4.75/hour. Subnet XX Celium on Bittensor does an identical offering, but you will only pay $1.50/hour. The future of AI is decentralised and distributed.
2.) Decentralised AI is doing things that were deemed impossible only a year ago. Subnet X does permissionless distributed training. Subnet X does permissionless distributed pre-training. Any part of the AI stack that you can find, I will show you a Bittensor subnet that does it faster, cheaper and more efficiently.
3.) Decentralised AI CANNOT work without a DLT. For a fuzzy logic consensus mechanism, you NEED a DLT. One way o looking at Bittensor is that it is a decentralised marketplace for machine intelligence. Peers rank and incentivize each other via Yuma Consensus, where validators score miners' outputs and distribute TAO rewards based on stake-weighted trust, preventing sybils with bond mechanisms and consensus thresholds.
4.) Buying TAO, you are buying XXX independent AI start ups. Just like how people buy the S&P500 or FTSE XXX. You are literally spreading your risk with TAO. You're not just stupidly yoloing into one single instrument. This isn't an AI crypto gimmick like FET or NEAR or ICP etc. This is hardcore AI that is literally on the bleeding edge of AI research. Just look at what all the subnets actually do:
Anyway, I'm on holiday, don't have time for a drawn out thread. Have a great day!
XXX engagements
Related Topics xai $reymi coins ai bittensor coins layer 1 coins dao coins depin coins oracle