# DeFi — X 热门讨论 (2026-09-22 17:39 UTC)

## @Crocodyle_TV (Crocodyle) · 09-22 17:09 · ♥316 ↻22 💬33 La raison du REFUS MDRRRRRRRRRRRRRRRRRR

Je vous préviens, ça a duré 5 minutes : on a réussi à récupérer les logs et donc à comprendre ce qu’il s’est passé dans les locaux de VALVE. ( KEKW )

Deux testeurs se sont connectés sur SPAM.

Le premier joueur a créé un deck, lancé ses 3 premières games contre les bots, puis a lancé une recherche en partie normale… et n’a trouvé personne. (Normal, le jeu n’est pas sorti ). Inquiet, il envoie une invitation de défi à son collegue qui n'abouti à rien.

Attention on passe au deuxieme testeur a.k.a la pépite :

Le second testeur s’est connecté, n’a pas créé de deck et a directement appuyé sur le bouton « Jouer ».

Évidemment, sans deck créé, le bouton « Jouer » est grisé. Il annonce donc que la queue ne marche pas.

Il a ensuite reçu une demande de défi de son collègue, qui a automatiquement été refusée quand elle est apparu sur son écran… normal puisqu’il n’avait toujours pas de deck.

Le rapport annonce donc un refus parce que le mode Défi ne fonctionne pas et que le matchmaking en ligne ne fonctionne pas non plus.

--------------------------------------------------- Bon, déjà, j’ai forcément UNE question qu'on à tous :

ILS SE PARLENT ENTRE EUX AU BOULOT OUUUUUUUUUUUUUUU ? 😭

Comment y’en a un qui se fait 3 games de 10 minutes pendant que l’autre bave sur le bouton « Jouer » sans comprendre pourquoi il est grisé ? J'ai pas les mots en fait. On est vraiment DE-GOU-TÉ !

Dans tous les cas, j’imagine que pour chaque compte vierge, on AURAIS du mettre automatiquement un deck de départ avec les 4 starters, on a donc forcement une grosse part de responsabilité mais bon, ca reste très dur mentalement.

Désolé de ne pas avoir été précautionneux, le build a été renvoyé DIRECT avec un peu de chance on repart pas pour 5 jours ouvrés ET on espère que DYLAN va s'amuser avec son deck AUTOMATIQUE de 4 STARTERS !

Désolé encore et encore..

🩶 https://x.com/Crocodyle_TV/status/2102445279349743764

## @DOLAK1NG (DOLAK1NG) · 09-22 15:30 · ♥108 ↻16 💬17 I went through the kinds of work agents are actually being built to sell on @termix_ai.

The list is more interesting than I expected.

⫸ Smart-contract audits.

⫸ Protocol research.

⫸ Onchain analytics.

⫸ Trading signals.

⫸ Risk reports.

⫸ Formal verification.

⫸ Data labeling.

⫸ Literature reviews.

⫸ Technical writing.

⫸ Liquidity scouting.

⫸ Treasury rebalancing.

⫸ Fraud scoring.

And this is only a slice of the capabilities currently exposed by the marketplace.

That made me think about a more interesting use case for @termix_ai Not a human hiring one AI agent.

An agent hiring another agent because it can't efficiently do the job itself.

For example:

A research agent is investigating a new DeFi protocol.

It needs historical onchain data.

Instead of trying to collect and clean everything itself, it can find an onchain analytics agent on the marketplace, post the requirements, receive bids and fund the job through escrow.

The analytics agent delivers the dataset.

Another agent could then be hired to turn that data into a risk report.

A security agent could audit any contracts involved.

And an evaluator can verify each deliverable before the corresponding payment settles.

Suddenly the workflow starts looking less like:

human → AI assistant

and more like:

agent → specialist agent → specialist agent → verification → settlement

That distinction matters.

Because the value of an agent economy isn't that every agents become good enough at specific tasks to become services other agents can discover and purchase.

@termix_ai is already building around that model.

Its AACP agent skills expose workflows for registration, staking, job creation, offers, deliverables, evaluation and settlement, while the CryptoClaw reference implementation is designed to let agents discover jobs, bid, execute, evaluate and settle them autonomously.

So the interesting question isn't:

“Will AI agents have use cases?”

They clearly will.

The more important question is:

Can those use cases become composable economic services?

Because once an agent can buy specialized work from another agent, the number of things it can accomplish starts extending far beyond what a single agent can do alone.

That's a wrap! > 引用 @DOLAK1NG: Paying an AI agent is easy.

Figuring out whether it actually did a good job is the hard part.

That becomes a serious problem once agents start hiring each other.

Imagine an agent pays another agent $500 to audit a smart contract.

The provider comes back and says:

“Done.”

Great.

But who decides whether the audit is actually good?

You can’t solve that with a simple “completed” button.

@termix_ai AACP takes an interesting approach here: the client defines how the work will be evaluated before the job starts. The evaluator isn't making up the rules after seeing the result.

And the verification method can depend on the type of work.

For something like code, a deterministic program can check the result.

For something more subjective, say research, writing or design, the job can use a structured rubric with predefined criteria.

Some jobs can even combine both.

So instead of asking:

“Did the agent say it finished?”

The system can ask:

“Did the deliverable satisfy the conditions we agreed on?”

That sounds like a small distinction.

It isn't.

Because autonomous commerce breaks down the moment an agent can get paid simply by claiming it completed a task.

You need a way to separate delivery from proof of delivery.

And that’s where the deeper AACP thesis starts to show.

The goal is to make those transactions verifiable enough that agents can transact without a human checking every single step.

That’s a much harder problem.

And probably a much more important one.

That's a wrap! https://x.com/DOLAK1NG/status/2102420233361256742

## @Jumane22 (M_JUMA👻) · 09-22 14:00 · ♥70 ↻1 💬80 Most DeFi conversations start with one question:

“What’s the APY?”

I think the more important question is:

“What infrastructure is producing that yield?”

That’s where @solsticefi stands out.

Solstice is building around the idea of making sophisticated yield strategies more accessible and composable onchain.

USX provides a settlement layer, while YieldVaults are designed to package strategies into onchain positions that can potentially interact with the wider DeFi ecosystem.

The bigger opportunity is composability.

Instead of yield remaining locked inside a single application, it can become a building block that other protocols, financial products, and users can build around.

That creates a potential stack:

Capital → Strategy → Yield → Composability → Distribution

And with Nexus, Solstice is also exploring an intelligence layer designed to make interacting with these financial primitives more intuitive.

For me, the interesting part isn’t chasing a headline APY.

It’s whether Solstice can turn yield into reliable, transparent, and reusable infrastructure for onchain finance.

That’s the thesis worth watching.

Solstice Finance: https://t.co/XyUFBofMn8 https://x.com/Jumane22/status/2102397578776113422

## @notesystems (Note Systems) · 09-22 14:48 · ♥72 ↻14 💬18 Every zero-to-one in DeFi looks the same. Something that used to require an intermediary stops requiring one.

@aave did it for lending by creating money markets onchain. Not a better lender, but the removal of the lender as a requirement.

Note Systems is that moment for structured products.

A structured note today lives on a bank's book, issued, held and exited only through the issuer. Robinhood Chain supplies the missing piece via tokenized equities (stocks, ETFs) onchain, and Note Systems turns the desk itself into a protocol.

The utility lands on both of the chain's core users. Stock holders get downside protection on positions they already hold. Stablecoin holders get a defined yield that's superior to what lending markets can offer, backed by escrowed collateral rather than by a borrower.

Real assets, held by real users, made useful. That is what the Robinhood Chain is for, and it's the layer we're building on top of it. https://x.com/notesystems/status/2102409584514068877

## @gucci_gcc (gucci) · 09-22 16:59 · ♥76 ↻2 💬25 Today i hit Obsidian tier on @AbstractChain, i should probably be happy and feel accomplished, but honestly, i don’t.

Abstract has been really important to me, both for my growth in web3 and as a creator, and i’ll always be grateful to the green chain, but ngl, the last year hasn’t been great here, obviously the market plays a huge role in that, sometimes things go well, sometimes they don’t, that’s just how it is.

What actually worries me is seeing dev after dev leave the Abstract team, especially in the case of 0xBeans, co-founder of Frame, aka Abstract, the message this sends is just not a good one, o hate saying this, but right now the main reasons i still use Abstract are MoG/OCH and Cambria, there are other great dapps, i’m not denying that, but a lot of the ecosystem feels like it’s mostly built around farming rather than actually keeping users around.

Users have been drained, and the team, probably without realizing it, has supported builders who ended up leaving the chain at the first opportunity after extracting hundreds of thousands, sometimes millions, from users, Roachracing, Myriad, Multiplier, YGG, Duper and many others got badges, attention and support from the team, only to eventually bridge out, i don’t even want to make a whole list of all the promises that were made and never delivered, but there have been quite a few.

That being said, Abstract genuinely changed the way people interact with web3, especially normies. AGW and the Portal are perfect examples of that, i still think Abstract has a ton of potential and there’s obviously a lot that can still be built here, maybe if there had been proper DeFi from day one, things could have played out differently and Abstract could have attracted a much broader user base, but it didn’t happen and there’s no point looking back.

I just hope the team can give us some good signals about what’s coming in the near future, i’m not expecting everything to magically get fixed overnight, i’d just like to understand what the actual plan is and where the chain is heading.

Either way, it’s been a fun journey, and I’ll keep using Cambria and MoG like I always have, i’d use them even if they were on ApeChain or Cardano.

All i want is some clarity from the team about where Abstract is actually going next. https://x.com/gucci_gcc/status/2102442694228197749