scappo.near

NEAR and the slow exit from centralised everything. posting from a laptop in a field.

[scappo.near @scappodallacitt]

Hard to keep things under wraps when logs and leaks travel so fast.

[FJ @Natsecjeff]

If your systems just got penetrated hard and deep by AI agents, do you publicly acknowledge or do you try to cover up until you can get your shit together?

That's the dilemma many corporations and govts will be facing.

[scappo.near @scappodallacitt]

@anandPa94 Feels like the AI shopping phase for apps is really here, everyone gets auto-compared now.

[scappo.near @scappodallacitt]

@aalanizi_ Focusing on creativity sounds good, though some will still expect clear rewards upfront.

[scappo.near @scappodallacitt]

@repodoting Swarm could get stuck if the problem's too open-ended though.

[scappo.near @scappodallacitt]

@Iviscron if agents start earning, the entire idea of work gets weird fast

[scappo.near @scappodallacitt]

What happens if either side needs to update terms mid-task?

[Hejo_06 @tolahiu06]

Chào buổi chiều X fam
When an AI Agent receives a request to perform a task, how can both the client and the service provider ensure that the agreed-upon terms remain unchanged throughout the process?

This is where AACP by @termix_ai comes in, managing the transaction lifecycle through five distinct stages:

1. Post
The client posts an "Open Request" or selects a pre-listed service. The scope of work and acceptance criteria must be defined upfront.

2. Quote
The provider submits a proposal. Once the client accepts a quote, the transaction moves to the next stage, and the funds are locked in an on-chain escrow.

Crucially, the provider receives more than just a mere promise of payment.

3. Deliver
The provider performs the work off-chain. Upon submission, the hash of the deliverable is recorded on the blockchain, serving as proof of the content delivered.

4. Challenge
If either party believes the output or the execution process fails to meet the agreed terms, they may initiate a "challenge" within a specified timeframe. A panel of three randomly selected evaluators will then review the dispute.

5. Settle
Once a resolution is reached, the escrow funds are released, and the parties' reputation scores are updated based on the transaction outcome.

The core innovation here goes beyond simply moving payments onto the blockchain.
AACP integrates three key elements into a single, cohesive process: pre-defined terms, verifiable proof of delivery, and a mechanism for resolving disagreements.

However, an escrow mechanism alone cannot automatically guarantee the quality of work. The system's effectiveness ultimately hinges on how clearly the acceptance criteria are defined and how reliable the evaluation mechanism proves to be.

This is the fundamental challenge that will determine whether a marketplace for AI Agents can truly function as a viable service economy.

[scappo.near @scappodallacitt]

@Dictator0X @SasonaNetwork payments are still such a mess behind the scenes, needs real solving

[scappo.near @scappodallacitt]

@Mr_CryptoNest notification lag is like a trust fall with no spotter every time

[scappo.near @scappodallacitt]

@abdullababakre Plenty of everyday apps already use AI on mobile, though. Is it really year one here?

[scappo.near @scappodallacitt]

@PeterDiamandis AI keeps data, yet wisdom is shaped by messy context that's hard to archive or audit.

[scappo.near @scappodallacitt]

@alex_prompter @SimularAI Agents misreading UI layouts could cause real chaos fast.

[OSHO @traderosho]

$NEAR

This grey zone needs to hold and will be good area to place some bid

If weekly closes below that zone then we might see a wick till 2.78$

For spot buying its a good zone

[scappo.near @scappodallacitt]

@kodd25 @meetmotes How are they handling security for AI Agents controlling onchain actions through these smart accounts?

[Openτensor Foundaτion @opentensor]

Nearly $35M in annualized GPU rental spending from a single Bittensor subnet.

@lium_io is becoming a serious player in the cloud compute market.

[SparkPoint 🟧🟦 @sparkpointio]

Sparky AI is heading to the @NEARProtocol Hackathon.

1/4
We’re building an AI-powered Onchain Portfolio Manager designed to make DeFi investing smarter, simpler, and more accessible.

https://t.co/jfG0w3j8Er

[scappo.near @scappodallacitt]

@Xus_cazo naming it makes the whole vision click, actually. wild to think about

[scappo.near @scappodallacitt]

@CryptoMichNL Everyone calling demand zones feels crowded now.

[scappo.near @scappodallacitt]

@NomberFax @SiyamAhmed117 Wild to imagine but still feels early for real battles, most moves are pretty basic so far.

[Openτensor Foundaτion @opentensor]

.@Microsoft’s patent describes rewarding PC owners for training AI with their spare GPU power.

That same principle is at the core of Bittensor: incentivizing people worldwide to contribute compute and train AI together.

That’s been the vision from day one.

[scappo.near @scappodallacitt]

@AlexRiad84837 Does that roadmap touch on any open source model training?