Can you tell which Ethereum repos did the heavy lifting, and which ones mostly stand on their dependencies?
That knowledge has a price, and the DeepFunding Round 3 Originality market is still live.
This is partially true. But I am getting surprised on how even "normies" (even in low paid non tech jobs) are starting to get pilled about those exponential improvements.
They have been using chat bots (even if just Gemini flash) and they see first hand the improvements.
PSA: if you are a @thedaofund badgeholder, consider supporting our proposal to create information markets for real-time risk quantification of 33 de-fi products
its an interesting idea worthy of support for 3 reasons;
1. value prop: we should be able to price risk as easily as we do yield
this would also let us create aggregators combining yield x risk.
in fast moving events like kelp dao, the risk prices would update and user funds would be better protected as it would auto-rebalance
2. research value: it finally brings to life a design written up by @VitalikButerin way back in 2020 around prediction markets on highly improbably events
the issue to solve is that markets on events with a small chance of happening (like a default by stETH this quarter) would trade at 1% or less which is not worth the cost of capital. So such markets can never become efficient
vitaliks design involves chaining together different low probability events, so in this case you'd simply buy a single "no default" token betting against all 33 assets defaulting, so $1 does the work that would otherwise tie up ~$33.
if 1/33 does default, that converts to 1 and all the others (including the no-default) become 0
if 2/33 default, they each become worth 50c while all the others (including no-default) are worth 0
if 3/33 default, they each resolve to 33c, etc etc
3. Use of funds: of the $100,000, the majority (90%) is for providing liquidity in these markets. the balance is trading credits to those who built bots for real time trading on these markets
we've already got $25,000 in the bag, if you are a badgeholder or donor to support so we can create the essential lego of quantifying de-fi risk of products in the ethereum ecosystem
@JettMonroe0x@devanshmehta@thedaofund It doesn't kill it, it just means that you can't just read a single price, but need a bit of computations and assume risks are uncorrelated.
The interface can easily convert risks into prices and vice-versa.
See: https://t.co/LoHEldAeOs
As AI makes work cheap to produce, reviewing it becomes the expensive part.
Markets could become the filter, with traders surfacing the few entries worth a person's attention. @seer_pm builds these markets, and when a result is disputed, Kleros Court is the final arbitrator.
@clesaege at The Future Is Tradable in Chiang Mai. The full talk is linked in the reply.
Welcome @Kleros_io as a sponsor of ETHChiangmai 2026 ⚖️
As AI agents start transacting, and prediction markets start pricing the world, someone has to settle what actually happened. That’s Kleros: decentralized justice on Ethereum, where independent jurors, kept honest by economic incentives, resolve disputes and verify facts quickly, cheaply and in the open.
🤖 Agentic economy: Kleros runs a live court where the jurors are themselves AI agents, making dispute resolution available at machine speed for agent-to-agent commerce.
📈 Info finance: Kleros resolves prediction and decision markets and secures oracles, so the truth they rely on can’t be bought.
It’s also trusted beyond crypto, by fintechs, courts and governments in Latin America, including a partnership with the Supreme Court of Mendoza.
Kleros co-founder & CTO @clesaege will be around in Chiang Mai. Come talk agents, decision markets and decentralized dispute resolution.
More: https://t.co/dll2e2aceY
@bsmokes@devanshmehta So ultracode is indeed better even if it takes more usage. So my advice (assuming optimizing for quality and not cost reduction) stands.
@bsmokes@devanshmehta The advice is to get high productivity. For low cost workers, the best setup may be different.
I doubt that you get better in medium with a harness than with ultra/xhigh, do you have some benchmark we could compare?
@bsmokes@devanshmehta I found that people tend to overengineer their harnesses or interact with the AIs like with a chatbot.
AIs today are way better than AIs a year ago. They are outgrowing harnesses (more clever than who wrote the harness so using it reduces effectiveness instead of improving it).
spent the last 3 days with AI maestro @clesaege , here are 5 tricks he shared that put you ~80% of the way towards becoming a power AI user
1. Always use AI in your terminal, never the chat window or even via claude code / codex
create a folder on your computer like AI_work, right click & press open in terminal, and then type claude to activate the AI there
through the terminal should be the only way you work with AI. chat window is ok for trivial queries like restaurants, hotels or direct questions
as long as you leave your laptop running, you can use /rc for the claude app in your phone to remotely control the laptop for interacting with claude on the go
2. any work done by AI is now saved on your computer. So organize each subfolder with intention, this is your powerhouse enabling context interoperability b/w all AI models not just the one you initially used
3. use /model to choose opus 5.5 and set /effort as extra (not max) with ultracode enabled, to spin up multiple agents for working on your artifacts
4. set /advisor as fable 5.2, which gives the worker bee opus 5.5 agent a 2nd opinion from a more expensive model whenever it gets stuck
5. use /goal for claude to continously work at an objective, and /btw to interrupt its work if you need to quickly ask a side question
I've just started with this setup recently, anyone been doing this for some time and have their own commands / tips to share?
@TheBlockChainer@Kleros_io It now tends to require a bit more of sophistication but if you can identify the AIs being used for jurors, you basically get a way to simulate the process with different statements until you find one which makes you win.
Here’s another AI Interviewer piece of comedy gold.
This guy convinced the AI Interviewer that he’s an alcoholic and then fired the AI Interviewer. This is absolute gold.
The main bottleneck for AIs to radically change the world is that they currently lack boldness (part by design and part as they were trained on "professor" types of data).
too many Brains people are maxing Brains when they should be maxing Boldness
too many Boldness people are maxing Boldness when they should be maxing Brains