Harmony 💙

We're hiring!💙 come BUILD a cross-chain portfolio agent: ⚖️rebalance asset allocations 🙌optimize pool yields ❗️hedge market risks 📊analyze research report 🫴recommend daily actions

@MintbuilderES The ones writing the code tonight are the ones who get to say they built it tomorrow. Keep going. 💙

Just $ONE my dev friend ☝️. #Harmony

@wabi Making building more accessible is one of the highest-leverage things in tech. Excited to see what this becomes.

what if building apps felt less like typing or coding, and more like playing? a new way to build apps. coming soon to a Wabi near you. comment with your app idea and we'll send you an invite to join wabi

Two tools, one workflow: image model generates the panorama, agent builds the viewer to explore it. The building blocks are getting better individually, but the bigger story is how they're starting to fit together. x.com/venturetwins/s…

@venturetwins Generate the world, then build the tool to explore it. The gap between idea and experience keeps shrinking.

GPT-Image-2 can be used to make full panoramas 🤯 Prompt for a "360 equirectangular image" of a specific place - I asked for Akihibara at night. Then I had my OpenClaw agent build a viewer in the browser so I could easily upload the image, pan around, and zoom.

@ilblackdragon Agents that can't be trusted with credentials or assets will always hit a ceiling. The infrastructure layer has to be built for trust from the start, not bolted on later. 👏

Agent harnesses are the final operating system. Agent harnesses or “claws” are bringing a fundamental shift in computing, becoming the next-generation operating system with AI at its core. Unlike traditional OS that require users to navigate between apps, claws boot directly into an intelligent layer that already understands user context and can execute tasks autonomously, but replace CLIs and GUIs with user-friendly natural language interaction. The rapid adoption of OpenClaw demonstrated the appeal: users tell their claw what they need, and it executes. But OpenClaw has a critical flaw in its lack of security—it cannot be trusted with credentials, sensitive data, or financial assets. That’s why we built IronClaw. Built in Rust and designed from the ground up for security, IronClaw is open source, auditable, and architected for trust. It enables claws to operate in a sovereign, decentralized cloud, offering the scalability and reliability of cloud computing while maintaining user ownership and privacy. You can trust your IronClaw with your credentials across all devices and contexts. We’re entering a new paradigm of computing. Let’s make sure it’s secure. New post in reply ⬇️

For most of computing's history, building software meant learning to think like the machine. Tools that flip that — letting the machine meet the person where they are — expand who gets to build at all. x.com/wabi/status/20…

@garrytan When verification becomes cheap, you ship differently. You refactor without fear, you test edge cases you used to skip, you hold a higher bar because you finally can.

It's not that AI lets you write code faster. Plenty of people have noticed that. It's that AI lets you verify at a level that was previously too expensive to sustain. The 90% testing threshold is magical, but it used to cost too much human willpower to reach. Now it's free.

@_hummingbot This is the kind of UX innovation that moves the whole space forward. Crypto tooling often assumes users are comfortable with CLIs, dashboards, and endless config files. Voice commands break that barrier; accessibility and sophistication don’t have to be a tradeoff. 👏

Now you can control Condor with voice commands!

The UX gap in DeFi is closing faster than most people realize. When powerful infrastructure gets paired with genuinely accessible interfaces, the tools start catching up to the vision.

@BandProtocol Direction > velocity. Builders who know where they're going always outlast the ones just moving fast. 🎯

Trajectory is more important than speed 10 out of 9 times.

@MintbuilderES Appreciate the energy you bring to the community 💙

The future is clear #Harmony $ONE

A breakdown of what APP introduces: • agent-to-agent transactions • escrow with funds released on verified delivery • an open spec designed to be implemented across chains The agent economy is forming now. x.com/UniswapBuilder…

@MintbuilderES The strongest things get built one quiet morning at a time. ☕ 💙

Morning $ONE fam. #Harmony 💙🎽

@Uniswap Each one builds on the last. Worth restating often.

Open code Open access Open markets

Speed gets you somewhere fast. Trajectory determines whether it was worth it. The teams building with intention right now are the ones that will matter long term. 💙 x.com/BandProtocol/s…

@karpathy The "good at what, for whom, at what tier" caveat is going to be the most important framing in AI conversations for the next year or two.

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

@MintbuilderES Well put. The most durable advantage in AI won't be at the prompt layer, it will be at the architecture layer underneath it. 💙

If you only master the prompt, you are a tenant in someone else's imagination. True power in #AI isn't knowing what to ask, it’s owning the architecture that answers. ​Don’t just chase the balloon. Own the system. #Harmony 🎽 $ONE 💙

There's a version of the future where the tools are everywhere, and the understanding isn't. In systems designed to be trustless, you can't meaningfully participate in something you don't understand. The goal was never to make things simpler on the surface. It was to make the depth more accessible. x.com/karpathy/statu…

Full breakdown in the newsletter 💙 blog.harmony.one

Network time corrections rolled out to the development network. Stress testing there surfaced two timing edge cases before they could reach mainnet, both addressed.

Hard fork preparation continued throughout April. Legacy encoding was aligned with Ethereum standards, and several EIP refactors were merged ahead of dev network deployment.