Pinned
People had a problem and thought:
‘Hey, let’s use distributed systems to solve this problem!’
Now they have 2^n problems…
Ivan
4,215 posts
Ivan
@allquantor
Joined October 2013
- 99% of so-called "Autonomous AI Agents" will go to zero, leaving you holding the bag! Why? Because most of them aren’t truly autonomous! Many outputs aren’t even generated by AI models—they’re human-written shills for worthless tokens. Let’s break this down. 🧵👇
- The $AR(weave) ecosystem is massively undervalued. It’s the only one offering the full-stack Web 4.0 experience: permanent storage for high-value assets, a DA layer, and ArFleet for faster, cheaper temporary storage—outshining Filecoin. With 7M Odysee users, there’s a huge chance
- The current cycle of AI Agents is a bubble. But every bubble also brings innovation! Data pipelines will be the technical breakthrough of this cycle—the ability to collect, process, and leverage crypto-relevant data effectively with LLMs. Let’s break it down. 👇
- Replying to @WillManidisI said it once and I’ll say it again: Celsius + Boxing + Cold shower + Modafinil
- Arweave fees are soaring as @aoTheComputer scales. At 1,200 $AR/day, that’s roughly 0.5% annual deflation; 9,000 $AR/day would hit 5% deflation. This ties protocol economics directly—unlike ETH, where L2s dilute value. Scalability beats scarcity. Study this.
- What a whirlwind week at @EthereumDenver few takeaways: - Finance? Practically a ghost town. DEXs and others with deep treasuries were conspicuously absent(officially as sponsors) hinting at the chilling effect of US regulations. - Witnessed an influx of 15+ L1/L2 projects.
- I want to address the recent concerns and, more importantly, share why we are more energized than ever about what’s ahead. The transition from Legacynet (Testnet) to Hyperbeam has been challenging because we are building something novel. Unlike traditional VC-funded projects
- Wild! We are testing competitive Autonomous Agents on @aoTheComputer, which are @autonomous_af stack. The effects are pretty wild. It might be a little early, but I doubt a scale like this is possible anywhere else.
- 🧵For the upcoming AO mainnet launch, here are some practical tips for writing more efficient, fast, and performant Lua code for blazing on-chain programs on AO – a mega thread 🚀
- If there will be a fully on-chain LLM with verifiable input/output and permanent + immutable model weights, I won't tell anyone -- but there will be signs.
- Transactions on Arweave over time stats.dataos.so/arweave?expand…
- Update on @aoTheComputer, been researching and talking with the team live, still seems very promising as a competitor for a world computer, I love the heterogenous nodes (any kind of hardware can join) and way more flexible computation limits than ICP. Still some concerns and
