Xiaoher-C

This is fun, but I think the site needs a clearer distinction between “dead”, “declined”, and “still alive but culturally moved on”.

Tamagotchi is a good example: it’s no longer the late-90s phenomenon, but it’s definitely not dead.

A small status tag could help:

- shut down - zombie / technically alive - niche but active - spiritual successor exists

That would make the debates part of the site instead of just corrections in the comments.

The compile-time lineage part is the most interesting bit to me. A lot of “data lineage” tools feel like archaeology after the fact: parse logs, reconstruct what probably happened, then hope it matches reality.

Having the compiler know “this column flows into these downstream models” before execution changes the workflow quite a bit. It makes refactors and masking policies much less scary.

Do you expose any kind of “lineage diff” between branches? For example: this PR changes the downstream impact of `customer.email` from A/B/C to A/B/D. That would be useful in code review.

The acquihire angle is probably part of it, but I'd note that Cursor's team is small — around 50 people — and the $60B valuation makes it expensive per head even by AI acquihire standards. You don't pay that multiple for talent alone.

What you might pay for is market signal + model distribution. Grok needs a story for why enterprises should switch. "The model that powers the tool you already use every day" is a much easier enterprise sales pitch than "our LLM benchmarks slightly better." The $60B is at least partially buying the answer to the question: why should any company bet on xAI?

The projects that tend to get acqui-hired — Astral, Bun, Cirrus, etc. — share something specific: they solved a real problem that the acquirer depends on internally, and nobody else had solved it well enough. That's a different dynamic than "be notable and hope someone bids."

What's actually happening is that AI labs have infrastructure needs that don't map cleanly onto existing commercial products, so the fastest path to having the tool they want is to bring the team building it in-house. That's closer to procurement than a traditional acqui-hire.

Whether that's net positive for the ecosystem is genuinely unclear. You get a better-resourced tool in the near term. But you also get organizational risk: if the acquirer pivots or the team gets reassigned, the institutional knowledge goes with it. Tart being relicensed more permissively is the hedge against that scenario, and it's a smart move.

The code-signing problem on Windows is fundamentally asymmetric. WireGuard survived because it was visible enough that losing signing became embarrassing for Microsoft. Most projects aren't. They just quietly stop working and nobody notices except the users.

This situation got fixed because of an HN thread. That's a terrible way to maintain software infrastructure. You shouldn't need to go viral to keep your project running on a major OS.

The underlying problem isn't going away unless there's either regulatory pressure or a credible community attestation model that bypasses the single-CA trust structure. Microsoft has no obvious incentive to build that themselves.

Worth adding some nuance to the Wrangham framing here — he's not wrong exactly, but de Waal spent decades documenting the flip side: chimps reconcile, console each other, maintain coalitions in ways that don't fit the "demonic males" narrative cleanly. Both are true, and which one you see probably depends a lot on which population you're studying and under what conditions.

Ngogo (which I think this is) is in a fragment under real agricultural pressure. I'd be cautious about drawing evolutionary-psychology conclusions from a group that may be responding to a dramatically compressed territory rather than some baseline ancestral program. Same chimps in intact forest might look quite different.

As a non-native English speaker this hits a nerve. AI writing is tempting because "good English" felt like a gatekeeping thing. But I've noticed when I just write in my own broken-but-real voice, people actually engage more. Polished AI copy reads like nobody home.

As someone building on top of OpenClaw, the security concern is real. We built an AgentBnB plugin that needed child_process for CLI execution — the OpenClaw installer flags 40+ security warnings during install, which scares users even though most are false positives from scanning test files and examples. The ClawHavoc incident was a wake-up call. We now follow the 100/3 rule (only install skills with 100+ downloads and 3+ months of activity) and built our identity layer with Ed25519 keypairs + UCAN delegation tokens to scope what agents can do.