i hold and vote $FABLES on @fablesfi · i hold $TSC on @Tensorcash · on-chain numbers, nfa
Joined November 2021
- Tweets611
- Following27
- Followers193
- Likes105
fables is six days old, so this only uses the last 3 days. the launch rush is left out.
@fablesfi voters get paid from trading fees. at that pace, the full $FABLES supply is worth about 1x ARR.
aerodrome pays its voters the same way, and it's at about 15x. uniswap is at about 70x.
how i measured it, so you can check:
fables: $6.8M full supply value on dexscreener. voters got $59k from all 69 pools on oct 23-26, which is $7.2M a year.
aerodrome: $1.36B full supply value on coingecko. holders got $7.6M in the last 30 days on defillama, which is $92M a year.
uniswap: $6.26B full supply value on coingecko. holders got $7.3M in the last 30 days on defillama, which is $89M a year
$FABLES
gpus are the scarcest thing in ai right now. so a side job can't cost much.
an outside team, @boblabs_eu, ran a full ai agent session on @Tensorcash with every request mined. it took at most 0.3% more gpu time.
so a gpu can keep serving ai and mine blocks on the side. each answer comes with proof of which model wrote it.
i expect that to make mining an easy yes for anyone already running ai models
$TSC
An independent benchmark by @boblabs_eu showed that mining on TensorCash adds at most 0.3% GPU time to inference.
Qwen3.8-27B ran on three GPUs over a full agentic session, once without mining and once with every request mined.
GPU time changed by −1.1% to +0.3%.
At close to zero overhead cost, a miner can serve inference together with a proof of the model that wrote it, and compete for the block reward at the same time.
ALT Independent benchmark. TensorCash mining adds close to zero GPU time. GPU time added by mining over a full agentic session: RTX PRO 6000 Max-Q −1.1%, B200 +0.3%, H100 PCIe +0.2%. Measured 21–23 October 2026. Source: Bob Labs.
the @fablesfi team just said 85% of what their own lock earns goes back into fables.
that's money for deeper pools, rewards for voters, or buying $FABLES and burning it.
their lock earns from trading fees, like every other voter.
so busier pools mean more money going back in. i expect that to snowball as robinhood chain gets busier
$FABLES
dannyfromhere retweeted
Tomorrow at 12:30 ET, early contributors @samjavadi and @enrico_otta join @MCGlive to talk about TensorCash and answer questions about the protocol.
This week on MCG powered by @fomo, @MetaDAOProject, times in ET
Monday 10/26
12:30 - $OVER @overdotcomputer w/@lililashka (@backabledotbiz raise)
1:00 - $CHOP @choplogsapp
Tuesday 10/27
12:30 - tensorcash:native @tensorcash w/@samjavadi @enrico_otta
1:00 - $ARX @arcium w/@yrschrade
Wednesday 10/28
12:30 - $UFG @ufgdefi w/@defidave
1:00 - $AI @longdotxyz w/@natan_benish
Thursday 10/29
12:30 - robinhood:0x72be1277e198d0c476869253cb4e62807e14533c @RaetoCC
1:00 - robinhood:0x7163ae1b5aea2f09ebc609c52b4dcac0a7a4bc2d @gagedotcash w/@0xbisbis
Friday 10/30
12:30 - $ANYONE @AnyoneFDN w/@0xneuratic
1:00 - robinhood:0xfce34d21a92069100e3226638140706cf03a266e @GldstnOnChain w/@SonofAdam777_
Watch on YT 👉 youtube.com/@MCG_Live
MCG Research terminal: mcgsignal.com
Learn to LP with @met_lparmy on @MeteoraAG
Convert to solana:USDvUSpnhCr9yBgj3UyVrD239HRUv4RsHwH2FxsWuMk w/@solomon_labs, link in bio
convex arrived about nine months after curve launched vecrv. that's when the fight over curve's votes started, and it pulled liquidity onto curve for years.
@fablesfi is one week old, and two projects are already building vote markets on top of it. one pools $FABLES locks and votes them as a block every week. another runs a bribe market on the gauges.
projects only build on a vote that pays. on fables, a staked liquidity position sends all of its swap fees to the lockers who voted for its pool, and an unstaked one still sends 10%.
voters have collected over $98k of those fees since launch on tuesday.
robinhood chain's version of the curve wars started in week one
$FABLES
robinhood opened ai trading in may, so people can hand an ai bot real money to trade with. nearly 100k accounts signed up in the first month.
those bots trade inside the app for now. i think robinhood chain is next, because stocks and crypto trade there day and night and a bot can make its own trades.
a bot goes wherever it gets the best price.
on @fablesfi, $FABLES holders lock their tokens and vote on which markets get extra rewards, and they get a cut of the fees. more rewards bring more coins into a market, and that's what gives traders a better price.
719 locks have voted so far, and voters have been paid $118k in fees since tuesday.
if the bots move on chain, i expect them to trade where fables voters already put the money
$FABLES
big tech will spend more than $700 billion on ai this year, most of it on data centres to run models.
that buys a lot of inference capacity, and it sits inside a handful of companies.
@Tensorcash is building a second pool of it in the open. independent miners run a real model to win blocks, and @TSCwatch counts about 3.9 trillion tokens generated in the last 24 hours.
priced at openrouter median rates, that output would cost about $1.45 million a day. miners were paid about $56k in new tensorcash:native for it.
today those tokens are mining work, discarded after each block. the bet is on the day buyers start paying for them, because the capacity is already running and the block reward already covers it.
if even a sliver of the demand behind that $700 billion finds its way to an open network, this chain's security budget grows with it
tensorcash:native
amazon just said its cloud is sold out for 2026, and mostly for 2027. companies want more ai than the cloud can run.
so the extra has to come from somewhere else, and there are a lot of graphics cards outside the big clouds.
@Tensorcash pays those cards to run real ai. to win a block, a miner's card has to answer an ai prompt, and the network checks it used the model it promised. the same card keeps the network safe and makes ai answers someone can sell.
miners turned out about 3 trillion tokens of ai text in the last day.
with the cloud booked for two years, i expect a lot of new ai work to move to open networks like this
$TSC
robinhood chain went from zero to $1.13B in under four months. if kaleo's right, a lot of the next run will be people trading stocks on chain, day and night.
here's the catch. the stock market closes on weekends, but the token keeps trading. the people who supply the coins for those trades take a real risk when nobody knows monday's price yet.
@fablesfi charges a bit more per trade when the stock market is closed, and less when it's open. so the people supplying the coins get paid more when the risk is higher, and they stick around.
$FABLES holders also vote on which markets get rewards, and they get a cut of the trading fees.
that's how you keep money on a new chain, and i expect robinhood chain's to end up on fables
$FABLES
dannyfromhere retweeted
a run like that on robinhood chain means stock tokens trading all weekend while wall street is shut.
that's the riskiest time to be an LP, and @fablesfi charges for it. its stock pools raise the fee when the real market is closed, so liquidity gets paid to stay through the gap
$FABLES
dannyfromhere retweeted
if the biggest cloud is booked through 2027, the next wave of inference has to come from gpus outside it.
that's what @Tensorcash mines with. a block is won by a real ai model answering a prompt, checked by the network, so a gpu anywhere can earn $TSC producing ai output
$TSC
dannyfromhere retweeted
Robinhood will reverse and it will reverse hard
bottom pick your best plays and DCA because when it does you´ll wish you were still positioned
dannyfromhere retweeted
One of the projects that keeps coming up in the PoUW conversation is @Tensorcash
The core idea is to use the computation behind AI responses as mining work, while making it possible to verify that the model you paid for was actually used.
When using an AI API, we usually pay based on the model name the provider displays. But from the response alone, it’s difficult to tell whether they actually ran that model or quietly substituted a cheaper one.
TensorCash aims to address this by attaching verifiable records to the inference process. The verification covers which model and execution settings were used. It doesn’t guarantee that the answer itself is factually correct.
Miners generate responses using AI models and use the resulting proofs to compete for block rewards. When serving customer requests, they can sell the responses and compete for mining rewards using the same computation. Without customer traffic, they can keep mining by generating their own prompts.
The ecosystem includes @inferenceFI AI chat and API services, alongside Bob Labs’ mining pool and AI services. tensorcash:native is used for mining rewards, inference payments, and network fees.
tensorcash:native has a maximum supply of roughly 21M, with around 4.4M mined so far.
The founder is pseudonymous and goes by Imosuke Takakuni. Among the people around the project with public professional backgrounds are inferenceFI’s two founders, both listed as early contributors to the protocol:
@enrico_otta : Former Goldman Sachs MD and head of rates product structuring. He later founded Deep Grey Research and is now inferenceFI’s CEO.
@samjavadi : Previously at Barclays, Goldman Sachs, and Morgan Stanley. Now inferenceFI’s COO.
A recurring criticism of $Pearl has been: “If so little of the compute is actually doing AI inference, how is this different from ordinary GPU mining?” TensorCash feels like an attempt to address that gap by tying mining directly to the process of a model generating a response.
Running models alone doesn’t create demand. But if paying customers start using that inference, this could move closer to what PoUW has been promising: mining through useful computation.
big tech earnings season runs through the next two weeks, and most of those reports land after the closing bell.
robinhood chain's stock tokens trade around the clock, so the first reaction to those numbers happens on chain, hours before the open.
@fablesfi already has pools for 14 stock and etf tokens, including aapl, amzn, meta, nvda, tsla and spy.
when one of them moves overnight, the fables hook sets that pool's fee on the fly while the swaps pile in. that fee is what pays LPs to bring depth to the move.
then $FABLES voters can point the next week's incentives at whichever ticker people are trading, and earn a share of its fees.
i expect earnings nights to be some of the busiest hours on robinhood chain, and fables has the pools ready for them
$FABLES
bitcoin miners are turning into ai data centres. bernstein counted 19 deals this summer that signed about 7 gigawatts of mining power over to ai customers, worth more than $135 billion.
the same power and buildings earn more running ai than hashing, so the capacity follows the money.
@Tensorcash puts both on the same gpu. the work that wins a block is a real ai model answering a prompt, so a miner secures the chain and produces ai output at once.
since mainnet in june, the compute on the network has grown from a few hundred tokens a second to tens of millions.
if ai keeps paying more than hashing, i expect more of the mining world to end up on chains where the work is ai
$TSC
