It'd be cool if you had a simple skill that I could install and run on a schedule to find high ROI tasks. I can imagine people signing up to perform these tasks with their slack utilization. On a schedule, the agent could wake up, check if its reset deadline is coming up or not, and assess its spare capacity vs the required work (ie, coming up against the 5h limit, or the weekly? How much of the allowance will go to waste and how much time does it have to complete this task?).
You can even imagine a second order agent with an outer loop assessing tasks, and smaller luna/sonnet level ones performing. You could even have the meta task of assessing task ROI as a task on the board!
Recently I wrote a blog post[0] expanding on a similar idea from the angle of ancient Greek philosophy, particularly Parmenides: to think is to compress.
Thanks! The hash chain is stored in Cloudflare D1 (edge SQLite). Each registration entry includes the previous entry's hash, so the chain is tamper-evident. You can walk it from genesis and verify every link via the public API (/v1/chain).
It's centralized storage with cryptographic integrity, not decentralized consensus. The tradeoff was deliberate: we wanted verifiable ordering and tamper-evidence without the overhead of a blockchain. Any client can independently validate the chain by re-hashing from entry 0.
Longer term, the chain data could be mirrored or anchored to a decentralized store (IPFS, Arweave, or even periodic Bitcoin/Ethereum anchoring) for stronger guarantees. But for now, the priority was getting the identity + reputation primitives right.
Very cool! Would love to check this out. Have been meaning to interview my aunts for years to write down my family history now that my dad is no longer with us. A tool like what you describe could be super helpful.
AngelList | Senior Software Engineers, Engineering Lead, Product Designer, Senior Data Engineer, etc | Full-time | Hybrid in San Francisco
AngelList's mission is to accelerate innovation by providing startups and investors with the tools they need to succeed. We have the world’s largest early-stage venture portfolio with $17B+ assets under management and $124B+ on our platform, delivering capital to over 12,000 startups. We work with some of the best emerging fund managers to ensure every deserving startup has access to capital.
I was hired into the role from one of these HN threads last summer. One of the things that attracted me most was the density of former founders and aspiring entrepreneurs on the team.
We're hiring full-stack engineers with experience building web applications at scale. Fintech knowledge is a plus, but not required.
I'm the hiring manager for our Funds Engineering team[0]. We’re building the data layer that powers our general ledger. We work on some of the most interesting datasets out there: We model capital flows in the venture ecosystem to understand where value accrues and, ultimately, which investors and founders win or lose.
I'd be happy to talk about these roles, about any open role we have at AngelList[1], or any questions you might about working here.
AngelList | Senior Software Engineers, Engineering Lead, Product Designer, Senior Data Engineer, etc | Full-time | Hybrid in San Francisco
AngelList's mission is to accelerate innovation by providing startups and investors with the tools they need to succeed. We have the world’s largest early-stage venture portfolio with $17B+ assets under management and $124B+ on our platform, delivering capital to over 12,000 startups. We work with some of the best emerging fund managers to ensure every deserving startup has access to capital.
I was hired into the role from one of these HN threads last summer. One of the things that attracted me most was the density of former founders and aspiring entrepreneurs on the team.
We're hiring full-stack engineers with experience building web applications at scale. Fintech knowledge is a plus, but not required.
I'm the hiring manager for two senior roles on our Funds Engineering team. We’re building the data layer that powers our general ledger. We work on some of the most interesting datasets out there: We model capital flows in the venture ecosystem to understand where value accrues and, ultimately, which investors and founders win or lose.
> Given React and Next’s wider use and longer history, it’s likely most LLMs are trained on more React and Next code than Svelte code. Ditto for FastHTML. This could lead to coding assistants being more effective when working with and suggesting code for established frameworks such as FastAPI, React, and Next.js.
Yes, but also more stale code from old versions which use patterns that the community has for various reasons moved on from. I ran into a lot of trouble with deprecated patterns while teaching myself react last year with assistants on the side. React 17 and prior version patterns kept coming up all the time.
You can even imagine a second order agent with an outer loop assessing tasks, and smaller luna/sonnet level ones performing. You could even have the meta task of assessing task ROI as a task on the board!
reply