Matt Devin

@LouisThibault87 @lobstermindset So the weight of owning something naturally shifts on whether it becomes valuable over time. So if you take out the financial aspect, you could say Web 3 is conceptually a flawed idea.

@LouisThibault87 @lobstermindset Fair. The emphasis on financial gains is somewhat tied to Web 3’s definition. The concept is to introduce “digital ownership”. But the open source nature of things means you don’t have traditional ownership model rights like proprietary exclusion, etc.

no hate to the people who left crypto, but i wonder if they're just going to run into the same problems in ai that they faced here. yes ai is more bubbly rn, but building real stuff still takes hard work and foresight and idk if the domain was the problem

@reisnertobias Your CEX’s second largest OI is ETH. If ETH didn’t “succeed”, your CEX would have a lot less revenue. Only crypto retards think a market place defines the market, not the product itself. But then again, this space rewards too many retards obfuscating this space’s purpose.

Hyperliquid built what Ethereum could have been. Reality is that if Ethereum succeeded we wouldn’t even need Hyperliquid

@LouisThibault87 @lobstermindset Relying on token be financially valuable with no clear economic reason to be so, is a sort of financial “degeneracy” or alchemy.

@LouisThibault87 @lobstermindset Even non-degenerate projects launch a token. You have DAOs, governance tokens etc. But the game theory for these token is, they are economic incentive/coordination devices, e.g. someone votes for protocol’s best interest to keep token valuable.

@lobstermindset I think too many crypto devs think money grows on trees and they are entitled to funding for their existence. It is just a bad assumption and terrible model to view how the world works. Economic gravity will hit them wherever they go.

@LouisThibault87 @lobstermindset Every crypto project launches a token, regardless the needs for it. How can devs blame the “degeneracy” when they are the ones complicit in propagating and benefiting from this focal point?

@lobstermindset I do have to admit that the finance-coded degeneracy is pretty specific to web3, and tends to be *extremely* off-putting to a lot of intelligent people. AI doesn't have that particular cultural problem.

@hellojintao @Jason Usually these VC types don’t bet on the public token directly. Their “alpha” is to fund “betas” in the ecosystem.

@hellojintao @Jason Maybe it is just a great place to launch a “AI startup” to eat TAO emission. VCs can just shove their struggling AI portco into the ecosystem subnet as an exit strategy via TAO emission.

how underwater is @Jason ?

@ryanconnor Shouldn’t the scale effect depend on the size of the ecosystem? Cosmos was a small cap ecosystem. How much size do you need to save a small cap ecosystem? And why was DYDX not big enough to save such a small cap ecosystem?

@MattDevin6 Think dydx was sub scale

Your favorite chain (Solana, Hyperliquid, Base, etc) has to win perps bc > perps are the most high volume & liquid instruments & > the liquidity network effects of perps will almost certainly pull activity and apps from adjacent ecosystems

@matar770 A lot of times, you just need to learn this space’s history and draw out the right comparisons to sniff out what is what.

@matar770 Hyperliquid token appreciation gave a wealth effect. DYDX token was just going down. Crypto “analysts” overanalyze “metas” and don’t acknowledge things are just driven by wealth effects.

Spent some time thinking about this... I honestly couldn't find an explanation.

@JoestarCrypto Before BTC, what “SoV” was ever capped? This SoV idea predates BTC. It is more about balancing supply against demand. Until crypto apps can show they can generate yield outside the ouroboros wheel, they are just second order effects. Second order things are riskier bets.

Lol as always this pissed off ETH believers I don’t get how you can see $ETH as a SoV when supply isn’t capped or think people will buy it just to grow economic security, delusional religion imo Reality is Ethereum would work the same even down 80% I love Ethereum, I really do, but not $ETH sorry Just look at $HYPE, accrue perfectly value with 16 validators lol, opposite of ETH

@JosephJacks_ @bittensor Often crypto VCs don’t realize they live in a bubble and pretend they can do better outside the bubble. You all don’t understand, these “idiots”, who got rich just because they are early, is how your “AI” L1s have such high valuation.

@JosephJacks_ @bittensor It is actually thanks to these “idiots”, your thing is above $1B MC. These “idiots” don’t care about details and chase “narratives” for pumps. If “financial mainstream” determines your pricing, don’t think it would be over $1B MC anymore.

The only thing holding back @Bittensor from being ranked as a top 5 cryptocurrency by market cap is that 95% of crypto holders are degenerate idiots. This changes the moment the financial mainstream realizes there are only two tokens that represent freedom from enslavement.

@MikeIppolito_ If you want app chains thesis to work, your first priority to find many overly successful apps existing L1s can’t handle, before worrying about building infra for them.

@MikeIppolito_ Why? Most VC funded apps can’t even find reasonable usage. Why do they even need a separate chain? And there is like only a handful of sectors ppl care about. Why over complicate what market wants?

The rollup idea wasn't wrong, we just don't need 20 gen purpose ones. The end state market structure is 2-3 vertically integrated gen purpose frameworks and many app or sector specific rollups.

@zacodil In that game theoretical/economic sense, it only seems interesting if these DePIN ones can produce better outcomes of existing free models, pressuring for-profit companies to open source more of their products to fend off competition.

@zacodil If these things can’t beat old gen models produced by for-profit companies, can’t these for-profit companies just release their outdated models over time to make these DePIN ones irrelevant?

Lots of TAO threads in my feed lately so I spent time digging into what Bittensor actually does vs what the narrative says. The headline achievement - training a 72B model across commodity GPUs with no central cluster. Sounds revolutionary until you ask: the result (Covenant-72B) barely edges out LLaMA-2-70B, a model from July 2023. You can download LLaMA-3-70B for free and it's significantly better. Massive distributed compute to produce something already obsolete. Cool proof of concept, not a product. The deeper problem is economics. TAO emits ~$365M/yr in token inflation against low single-digit millions in actual subnet revenue. Chutes is "85% cheaper than AWS" because emissions subsidize compute, not efficiency. When halvings keep cutting rewards, that pricing breaks. Now look at NEAR (~$1.7B mcap) - founded by Illia Polosukhin, co-author of "Attention Is All You Need," the paper that started the entire transformer/LLM revolution. The AI stack is serious: DCML (Decentralized Confidential Machine Learning), Private Inference, IronClaw for confidential compute, live AI agents on market(.)near(.)ai already doing real jobs. Confidential Intents launched March 1 - private cross-chain execution via ZK + MPC + TEEs. This is AI infra built for agents that need privacy and verifiability, not just raw GPU hours. And unlike TAO, it actually generates revenue: → $14.4B cumulative Intents volume across 35+ chains → Fee switch live - ALL fees convert to $NEAR buybacks → 2.1M NEAR bought back in 3 weeks → P/S ~28x on ~$53M annualized fees vs TAO's effectively infinite P/S → Deflationary threshold at $177M daily vol, currently $77M growing 17% MoM Both building AI infra. One subsidizes it with inflation, the other monetizes it with real fees. That gap won't stay invisible forever.