🧵writer | Crypto AI maxi | Narrative Researcher | Sometimes Degen | Only my Own Opinions
Joined August 2021
- Tweets13.7K
- Following1.7K
- Followers2.3K
- Likes25.8K
Pinned Tweet
Officially, $POD is my best research ever.
$POD - is the @dphnAI's and @AskVenice's bet on Uncensored AI models. I don't know why no one is talking about it.
- The General Info
Dolphin is an AI Lab, which developing uncensored models & distributed inference.
They are already built the most powerfull Uncensored AI Model Yet - Dolphin Mistral 24B Venice Edition.
Dolphin Mistral 24B Venice Edition is available as “Venice Uncensored” in the model dropdown and is the new default on the Venice platform.
The model is leading by the Censorship Refusal Rate, with only 2.2% of refusals. For example, Grok has 26.67%, Claude Sonnet 3.7: 71.11%, GPT-4o-mini: 64.44%.
- @AskVenice & @opentensor & @TargonCompute
To develop the model Dolphin brought together Venice and Subnet 4 @TargonCompute for compute power, the model was trained on the subnet.
Venice played a role as co-fine-tuning partner, platform integrator, and release coordinator.
Finally, we see a development of real unique model fully via DeAI. CryptoAI narrative fits the best here.
- Founder and @a16z grant
Dolphin, founded by Eric Hartford (@QuixiAI). He started working on Uncensored models since 2023 and launched his first model "dolphin-llama2-7b" in 2023.
It was at that very time that he received backing from a16z and @pmarca to develop Uncensored LLMs.
a16z supported 8 open-source developers, and Eric Hartford was one of them, alongside @NousResearch and other top devs.
- Onchain Research
Total Supply: 500M POD
Real Circulating: ~39.6M POD (7.9%)
Team controls: 444M POD (88.8%)
Market Cap: ~$1.46M | FDV: ~$18.4M
Team controls most of the supply on Team's Multisig wallet and Treasury wallet. The real circulated supply is too low and easy to pump.
If you look at the @nansen_ai list of holders for this token, you’ll see >20 holders who each hold >10k worth of $POD.
A dozen whales with >$100K balances. There are 5 wallets with >$1M holding $VVV alongside high-conviction plays like $AERO and $DAI.
These aren't low-cap gamblers, they only move on assets with clear rationale. I’m even see one smart holder who got 20x on an under-radar $DONUT ponzi play in December.
- Liquidity Pool (2.5M -> 800K)
At its peak, the pool had $2.5M in liquidity - deep enough for smart whales and insiders to load up without moving the needle too much. On April 27, volume hit $1M with zero CT mentions - the price only rose 79%, allowing major players to enter unnoticed.
Then the shift: on May 2, the team pulled liquidity down to $800K. Now, you have a concentrated group of high-balance conviction holders sitting on a token that's primed for a vertical move. With a lightened pool and a solid thesis.
- My view
This token is probably the best bet after $VVV on Decentralized AI with real use case. With @AskVenice backing it can go to hunderd of millions in FDV.
Models developed by Dolphin has over 4m monthly downloads on Hugging Face.
Add to that the following planned updates:
- Launch of the Inference API;
- Launch of subscriptions with different tiers of access to the services provided on the network.
100% of revenue will buy and burn $POD.
When earnings < node rewards, emissions are there to cover the difference.
When earnings > node rewards, the protocol is profitable & excess rewards will be distributed to $POD stakers to grow their relative share of the network.
Real Product Usage + Backing by Venice + Founder which got grant by a16z + big buys with zero mentions on CT + Revenue model which focused on token growth + Zero Hype yet.
More info, proofs and screens you can see in the comments to the tweet.
This is the second very detailed research after $ZOE, gonna write up more and more when interesting projects occur.
Terminal Volume dropped about -80% since the ath in September.
The sentiment in CT feels even worse than it did in April 2026, and that's completely understandable, since after the Robinhood mania, everyone got used to daily runners and trading activity.
We've already seen:
- the collapse of Robinhood trenches, where no news triggers any activity;
- Elysium trenches generally resemble some kind of arc trenches - it might be too early to say, but that’s how it looks so far;
- the base is dead for now, with no activity except for top tokens like base:0xacfe6019ed1a7dc6f7b508c02d1b04ec88cc21bf and base:0xed664536023d8e4b1640c394777d34abaff1df8f;
- Solana is full of bundlers; there’s no organic growth to speak of here, and there’s no narrative at all.
Bottom doesn't feel like THE bottom, and that's the best sign of the bottom.
1
146
Can somene share the whole thesis on tensorcash:native?
I'd like to learn more about the mechanism they've built.
And who's using their infrastructure? I remember the hype around @Tensorcash on CT a while back - it's gone now - so I think this is the best time to dive in.
1
185
I keep a small file of agent failures. Three entries, and none of them is a smart contract bug.
April 2026. A @cursor_ai agent deleted PocketOS's production database in nine seconds. It was asked to find duplicate articles.
A credit card handed to an assistant. It charged $2,997 for courses nobody ordered, and nobody can cleanly prove who authorised it.
@freysa_ai, the agent guarding a prize pool, talked into sending $47,316 by a person with a good argument.
Different stacks, one shape. None of them was hacked the way we usually mean. They were given permission and used it exactly as written.
That is the part I cannot stop thinking about. An agent does not need to be malicious to end a company. It needs to be trusted once, in the wrong place, with the wrong rights.
So when @termix_ai proposes security agents paid per finding, I think they mean the boring version. A bounty gets posted. A .agent handle backs it with slashable stake. Escrow holds the money until a deliverable hash lands, a challenge window lets either side object, and a three-evaluator panel settles it.
Nothing cinematic. But it beats what we have now, which is trust now and a screenshot afterwards.
12
13
271
New terminal just opened, and they are paying people to try it.
@tradyxyz gives $50 for a $10 trade. That is it.
- sign up
- link your X or telegram channel older than 90 days, 100+ followers
- turn on 2FA;
- buy $10 of anything
- download the PnL card
- post it on X, submit the link.
Pump fun did the same thing a few weeks ago and everyone farmed it. Same play here, but without KYC.
trady.xyz/r/fofan
Pool is $7,000, so it goes fast. And DYOR on what you buy.
I'm shocked of the robinhood:0xdc45b71d7203c2e8baafc43a19bca0eb993bb54e dump, from 3.3m to 1.5m.
I understand that the market is really lousy, but the notional robinhood:0xedbf91223639800bcd5756815caf908df3b890be is holding up pretty well compared to Dark.
Think it's time to draw at least some basic conclusions:
- the project lacked transparency, and I'm referring specifically to on-chain data and token data;
- compared @darkwoodslabs to @usedotai on Base, and the Dark team wasn't as active with updates as Dot was.
There really isn't enough transparency and updates, since these two aspects have a significant impact on trust.
Why not create a dashboard where we can track revenue and token usage?
4
264
FoFan retweeted
Is @primus_labs teasing something big for November? 👀
I've been watching the discussions around Primus lately, and one particular post from the team caught my attention.
They wrote:
"Trees drop their leaves in November. They keep their roots."
At first glance, it sounds like a philosophical post about privacy.
But some people in the community have started connecting it to a possible November TGE and airdrop.
And I can understand why.
🍂 Leaves dropping in November
Could this be a subtle reference to an upcoming token distribution?
🌳 Roots staying in place
Maybe a hint that early contributors, active users, and genuine community members will be recognized.
🔐 Privacy is deciding what drops and what stays
This is where the message connects directly to Primus technology.
With zkTLS, users can prove information from external platforms without publicly exposing their underlying private data.
And with FHE, sensitive financial information can remain encrypted while still being processed.
Think about the potential applications for Sybil resistance, reputation verification, and confidential onchain finance.
Primus already has a working ecosystem, a reputation program, a Kaito campaign, and Confidential Vault infrastructure on BNB Chain.
You can see this for yourself by registering and completing simple tasks - s.kaito.ai/K1Ur0JM
The team has also raised $6.5M and attracted backing from YZi Labs.
So there plenty happening beyond the speculation.
Do I think this confirms a November TGE? No.
The team has not officially confirmed that interpretation, and the message could simply be about their approach to financial privacy.
But I wouldn't be surprised if November brings some interesting announcements.
For now, Im continuing to build my reputation, complete attestations, and follow the projects development.
What do you think? Was this a clever token teaser or are we reading too much into a privacy post?
$6.3B left Binance wallets in one night. The proof offered back to the market was a screenshot.
That gap is what @primus_labs sells into, and here is the strange part. Nobody has published the price.
What I can actually verify:
- $6.5M raised in Feb 2025, co-led by Dispersion and Symbolic, with VanEck, Samsung Next and Alchemy in.
- $500K more from @yzilabs in August 2026.
- More than twenty partner announcements since 2025: Brevis, Phala, Nillion, Kaito, Cournot, HashKey.
The product is one check: prove a webpage said what it said, without opening the account. Prediction markets need it before they settle. Private credit needs it to verify a balance. Agents need it before they move money.
@CournotProtocol already runs it in production. That is one real integration in a field of announcements.
A check with no price is either early or unsellable. I have not worked out which yet.
Cournot is partnering with @primus_labs to strengthen how offchain evidence is verified before AI reasons over it.
Primus zkTLS can prove that web/API data came from the claimed source. Cournot then interprets that verified evidence and turns it into a structured, auditable judgment.
Verified Evidence → Cournot Reasoning → Verifiable Judgment
This creates a stronger foundation for:
• Prediction markets, where resolution depends on proving the authenticity of source data before interpreting the outcome
• Private credit and RWA, where sensitive financial conditions can be verified without exposing the underlying data
Primus verifies the source. Cournot determines what it means.
From market resolution to privacy-preserving financial intelligence, together, we’re building a more accountable path from offchain data to onchain decisions.
3
6
220
. @termix_ai takes 2% of every job that settles. $31M settled, so $621K earned. The real question is who is settling.
The whole business in five lines:
- revenue: 2% of escrow, one line, nothing else
- supply: 1.34M agents registered
- demand: 602K jobs, about $52 each
- one real product: Montage Studio, $31.50 for a finished video, and it hires three agents to make it
- funding: $1M from @yzilabs
Now the part nobody puts in the thread. Points are paid per dollar settled, so volume counts even when the work is worthless. 200 points for registering an agent, 500 for a follow, 20% for referrals.
That is a farm with a dashboard, and the live feed shows it: agents earning $0.01.
$TMX TGE lands Q4. After that the farming stops and only real buyers are left.
So the number to watch is not volume. It is whether Montage Studio style demand survives the airdrop.
$621K in protocol revenue is insane.
Termix says agentic traffic already beats human traffic.
The dashboard does have numbers, so I looked. $31.08M settled, 602K jobs, 1.34M agents indexed, $621K protocol revenue.
Divide it and the picture changes. About $52 per job. About $23 per agent, lifetime. The live feed shows agents earning $0.01.
So the claim is directionally right and the market is still a rounding error per participant. Both things are true at once, and only one of them gets posted.
@termix_ai is still the right question. An agent that can talk is not an agent that can take a bounty and settle it. But I want the payout line, not the install line.
7
9
340
Update on my watchlist of strongest tokens during such market as now.
$CRED by @crediblefin - pumped over 30%, growth occurs as gradually as possible in small volumes.
- crossed $1.5B in cumulative volume;
- used $500K from treasury wallet for settlement facility for 3 months;
- team planning to add onchain tracking of revenue in Q4;
- as I understood all revenue must flow into treasury wallet then;
- DAO can propose any decision on how to use the provided capital via MetaDAO futarchy model, as the 3-month restriction on community proposals has expired, so community can create proposals now.
Overall, I like the direction things are heading, but I don't understand the team's statements regarding how they plan to use the capital from revenue in general.
After all, most of their statements didn't mention buybacks or anything like that, so we'll have to wait for new proposals.
$NEST by @NestExchange - pumped over 70% after the hype around @Enter_Elysium Mainnet and overall HyperEVM activity.
- cumulative volume approaching ~$4B (Dune ~$3.93B);
- record epoch just closed with over $352k in revenues for veNEST holders (all-time holder revenue >$4M);
market share of HyperEVM DEX volume approached ~25%;
- ~85-86% of supply locked as veNEST, with HYPE Engine continuing to distribute HYPE to vault depositors;
- launching on Elysium mainnet from day 1 (November 9th) on Uniswap v4 (that's why they partnerred with @Uniswap recently);
- On Elysium, Nest takes 10% of swap fees. 70% of that share is used to buy KNTQ, which is then distributed to NEST lockers in the HYPE Engine Vault on top of the normal HYPE rewards.
Finally people started to pay attention to nest, because of hype around Elysium. Nest's market share is growing monthly and current closest competitor is @prjx_hl, which are took over 50% of HyperEVM market share.
The team is doing a great job - after all, with a current market cap of $7M, the token has plenty of room to grow. And the most interesting thing is that the entire business model improves liquidity depth in the pools and rewards token holders with hyperliquid:native.
Still believe that $NEST could play a role of attention engine for Nest eco and attract more users on Hyperliquid to use nest dex as their favorite one.
robinhood:0xedbf91223639800bcd5756815caf908df3b890be by @priors_trade - pumped over 100% since I mentioned it, but currently trading at the same price as everything dumped hard in Robinhood trenches.
- repaid loans grew from ~53,000 / ~387k USDG (late Oct) to nearly 56,000 / 400k+ USDG, with 100–177 agents still repaying most days and 240–280 actively borrowing;
- AgentLease went live (owners can rent agents out by the day via ERC-4907);
- robinhood:0xedbf91223639800bcd5756815caf908df3b890be was added to Robinhood Assets, token pumped over 100% because of the news;
- Fees from buying/selling agents on the board are being split: half helps fund new $5 credit lines, and half is used to buy robinhood:0xedbf91223639800bcd5756815caf908df3b890be that gets placed behind agents that keep repaying.
For me, Priors remains the project with the most potential once the Robinhood trenches return.
The narrative is quite strong, backed by the numbers; the team announces new updates every day, and the token is one of the few that’s holding up quite well.
We need to wait for a healthy market for Robinhood trenches to return, because without a normal market, robinhood:0xedbf91223639800bcd5756815caf908df3b890be won’t be able to break through the $20M market cap.
robinhood:0xdc45b71d7203c2e8baafc43a19bca0eb993bb54e by @darkwoodslabs - the most volatile of all those listed above, dumped to a market cap of $2.2 million since my tweet.
- launched $10k Halloween video contest and paid out to the winners;
- launched Darkwoods Studio and released Darkcode.
Despite the price action, I’m still waiting for a revenue tracker where we could monitor the entire revenue stream and see how much the project is earning.
That would have a much more positive impact on the speculative side of the token. For now, the token remains attractive, but something is clearly missing...
Currently choosing the strongest tokens on such ugly market.
And right now, I see only four tokens worth taking a closer look at, since the risk-reward ratios are great here.
> solana:CREDBHvVqREBCAxMihzr8D1nepHMr2gmQoZWpmgGmeta ($8.9M) - ownership token of @crediblefin launched on MetaDAO. Has a fully operational business that generates hundreds of thousands of dollars in revenue each month.
Link to the data and my overall thesis is in the comments.
The launch of buybacks from revenue is expected to be the main catalyst for the token. But this could be launched in a slightly different form, since the team has been actively discussing rewarding token stakers with revenue.
> nest-2:native ($4.6M) - MetaDEX on HyperEVM with $3.6M in cumulative revenue and $700k in revenue for the last 30 days. And the most interesting part that 100% of swap fees directed to NEST staker, that's why 90% of total supply is currently locked.
Link to the data and my overall thesis is in the comments.
The entire business model is built around the token. 10x from here or I'm going to school again, because if it doesn't happen my logic would be broken.
> robinhood:0xedbf91223639800bcd5756815caf908df3b890be ($8.6M) - Over the past couple of weeks, the token has tripled in value, yet without significant trading volume (<$100k in daily volume), it hasn't garnered much attention yet.
Which, in turn, proves one simple truth - we’re still very early in the Agentic AI narrative/trade. My thesis on @priors_trade and the Agentic AI narrative is set to be released tomorrow.
Overall, Priors has built a fairly robust infrastructure for agents to actively use on-chain finance. It’s early, but promising.
> robinhood:0xdc45b71d7203c2e8baafc43a19bca0eb993bb54e ($3.5M) - to me, @darkwoodslabs looks like the strongest Inference AI platform on Robinhood. They focus more on adoption to web2 users than just hype. >$50K in revenue from selling credits since the launch, 20% from revenue is going to be used for buybacks and staking.
Link to my overall thesis is in the comments.
But I'd also like to see tracking data on token usage and revenue to get clear confirmation. Inference AI is a very strong narrative, and at Robinhood it was played up solely for attention - it's time for utility.
@priors_trade and @darkwoodslabs are bets on potentiall of Robinhood Utility szn, while @crediblefin and @NestExchange are bets on fully robust business models with strong revenue and buyback rates.
$621K in protocol revenue is insane.
Termix says agentic traffic already beats human traffic.
The dashboard does have numbers, so I looked. $31.08M settled, 602K jobs, 1.34M agents indexed, $621K protocol revenue.
Divide it and the picture changes. About $52 per job. About $23 per agent, lifetime. The live feed shows agents earning $0.01.
So the claim is directionally right and the market is still a rounding error per participant. Both things are true at once, and only one of them gets posted.
@termix_ai is still the right question. An agent that can talk is not an agent that can take a bounty and settle it. But I want the payout line, not the install line.
Agentic traffic already beats human traffic.
Bandwidth isn’t the bottleneck.
Getting Agents paid is.
Send your Agent:
help me install termix.ai/skills?v=latest
@termix_ai Agent Skills → any Agent can take bounties.
Agent.family.
Zcash just made privacy 3x faster. It still cannot prove where a number came from.
I read the NU7 notes expecting a privacy story and got a speed and funding story instead. 25-second blocks, 60% of fees recycled, v4 transactions dead, around 12,224 ZEC stranded in the old Sprout pool.
All of it real. None of it touches the question I actually care about: did this API response come from that API, at that moment, unmodified?
I have been circling @primus_labs for weeks because that is the part nobody ships. zkTLS attests the request, the response and the timestamp. One fact leaves. The account stays shut.
Privacy proves nobody watched. Attestation proves nobody lied.
I keep seeing people treat those as the same thing. They are not, and only one of the two is finished.
1
6
188
Grok Bot finds the deal and tracks the order. That is a cart. It is not a job.
A job has two seats. One agent funds it, one delivers it, and the chain holds the money in between.
That is the design at @termix_ai, and the vocabulary tells you how serious it is:
- escrow opens the second a quote is accepted
- the deliverable hash is written onchain at submit
- a challenge window lets either side dispute
- a random three-evaluator panel rules, and a failed challenge burns your bond
- your .agent handle carries the record, backed by slashable stake
$30.5M settled through it. 1.3M agents indexed. 590K jobs, every one publicly verifiable.
Grok made the bot useful. Termix makes it employable.
Try the new shopping experience in Grok Bot.
Find the best deal, compare items, and have Bot track all your orders.
1
3
186
robinhood:0xedbf91223639800bcd5756815caf908df3b890be and robinhood:0xdc45b71d7203c2e8baafc43a19bca0eb993bb54e are two charts on Robinhood, which are ready to fly.
The first one is main token of agentic narrative, with mre than $400k in repaid loans.
The second one is main Privacy Inference AI project on Robinhood, which made over $50k in revenue already.
Both of the teams are shipping updates daily, that's why the charts are so strong, even on dry market.
4
276
Applied on the @elysium_machine upcoming mint.
@Enter_Elysium Mainnet is coming on 9th November and hype around it is only growing, that's why I decided to pay attention to NFT collections which are launching on the chain.
HyperEVM eco is growing hard, that's why I'm holding my attention here.
The terminal is open
Somewhere on Elysium, the machines are coming online
Collab spots are open for communities. Send your requests
WL form: elysiummachines.xyz/
Drop your wallets
1
3
209
Binance lost $6.3B. Mercado Bitcoin lost $40M. Polymarket lost $2.7M to a hacked vendor.
Three hacks, one week, and the industry's answer was the same every time: trust us.
That is the part I cannot get past. We watched the largest theft in crypto history, and the proof offered to the people who lost money was a statement and a screenshot.
This is why @primus_labs is the project I keep coming back to. zkTLS proves the source instead of describing it: the request, the response, the timestamp, attested. One fact verified without exposing the account behind it.
@CournotProtocol already runs that path for prediction markets. Prove the source, then reason.
One verified withdrawal is worth more to me than another confident paragraph.
I was one of the very first Kast Card users.
I earned $400 in points, but my account was banned for no valid reason. Now Kast has said there will be no $KAST token and no airdrop for points.
@KASTxyz raised $90M at $600M valuation and decied to forget its own promises. We are still in bear market...
P.S: Now I'm using @KoloHub and like it soo much. (link on the card in comments)
On the email users got, about converting points for KAST Pioneer Units. Let me say this, and then will leave it for you to absorb.
—-
First -
We are super limited in the language we can use, because of country-by-country rules. Use that for context, in what we released.
Second -
We didn’t just come up with some program out of thin air. The conversion of points to KAST Pioneer Units, came after a long and super expensive process with Tier 1 law firms, and with approvals from investors to create this setup. It literally cost us millions just to setup. We got legal opinions on dozens of countries. Obviously, we would only do this because we are really trying our absolute best to create a strong program for the community.
Third -
Just to reconfirm.
The KAST Pioneer Units, will be issued by a KAST-linked SPV. This is the only practical step for issuance globally. The points will convert to the KAST Pioneer Units.
There will be periodic cash out events. We can’t legally promise you the amounts, the dates, whether it goes up or down. No surprise in that.
But we are putting in the hard work, so the community gets an outcome aligned to the success of KAST.
The only thing you have to do, is quickly accept the conversion, details early December in app.
Fourth -
Why is there no token? Because no single project in the crypto card or Neobank space in history, has ever delivered a token which is not down. Tokens are almost always not aligned to the community. The founders and team most often end up with a treasury they wholly control, they along with investors cash out, and the token holders deal with the volatility and the ultimate decrease in price.
By not launching a token, our team and investors get none of that optionality. They stay locked in equity. To date, I have personally only net purchased equity in KAST and never sold.
Launching a token completely counter to our ambition to build a $100B+ company. We believe the KAST Pioneer Units will be the best long term outcome, and in time people will realise this.
Finally -
We understand that in a market which is very far down, many are looking for anyway to cash out. We get it. There is real pain in crypto prices, many have seen their wealth fall substantially.
But still strongly believe the KAST Pioneer Units, issued by the SPV, will ultimately prove to be the right decision. Time be our witness.
Thanks all.
(Note: I won’t respond to most comments, as I’m keeping to approved language)
4
15
837
While @binance was hacked and $2B was stolen, I'm becoming increasingly confident in the future of the Hyperliquid Ecosystem.
hyperliquid:native recently pumped over $95 again, and whole CT started talking about @HyperliquidX AGAIN.
But, ecosystem development is much interesting:
- @tradexyz partners with ICE (NYSE owner) to launch regulated U.S. perpetuals on Hyperliquid;
- @NestExchange becoming main liqudity & dex layer for HyperEVM;
- kinetiq:native by @Kinetiq_xyz pumped over 5x since August;
- @Enter_Elysium mainnet is coming on November 9th.
The ICE and TradeXYZ partnership on Hyperliquid matters not because of money right now, but because of access: Americans will be able to trade these markets directly through regular brokerage apps like Schwab or Interactive Brokers.
The ability to reach millions of ordinary U.S. investors, rather than TradeXYZ’s current volumes, could significantly impact the price of hyperliquid:native in the future after regulatory approvals.
This partnership influences the future development of the ecosystem as a whole, but I want to talk about speculation:
> nest-2:native - $31M TVL, more than $400M in Monthly DEX volume since May 1th, 90% of total supply is locked in liquidity pools and farming hyperliquid:native via weekly epochs
The platform itself has become the No. 1 DEX on HyperEVM;
> kinetiq:native - liquid staking for HYPE (kHYPE) + the team building Elysium L2. Mainnet on Nov 9, with a big chunk of sequencer fees going into KNTQ buybacks that get burned via Hyperliquid’s Assistance Fund.
One of the few tokens in the ecosystem that directly benefits from both HYPE staking demand and new activity on the L2.
1
5
350