CactusBlue

Mechanize | San Francisco, CA (Onsite) | Junior SWE ($300k + equity + bonus), SWE ($350k + equity + bonus), Research Engineer, Alignment ($300k-$400k + equity + bonus)

Applying takes <2 min: https://www.mechanize.work/apply/

We build environments that frontier AI labs use to train their models to do real-world software engineering. We have more demand than we can keep up with, so we're hiring fast.

Mechanize | San Francisco, CA (Onsite) | Junior SWE ($300k + equity + bonus), SWE ($350k + equity + bonus), Research Engineer, Alignment ($300k-$400k + equity + bonus)

Applying takes <2 min: https://www.mechanize.work/apply/

We build environments that frontier AI labs use to train their models to do real-world software engineering. We have more demand than we can keep up with, so we're hiring fast.

Mechanize | San Francisco, CA (Onsite) | Junior SWE ($300k + equity + bonus), SWE ($350k + equity + bonus), Research Engineer, Alignment ($300k-$400k + equity + bonus)

Applying takes <2 min: https://www.mechanize.work/apply/

We build environments that frontier AI labs use to train their models to do real-world software engineering. We have more demand than we can keep up with, so we're hiring fast.

How do you know that this isn't already priced by quant firms?

is privacy of financial transactions not a valuable problem to be solved in your opinions?

If you have a big enough portfolio, direct indexing (using something like Frec or Wealthfront) could be an interesting option, and weighting the companies that you don't want at 0.

It feels like the Microsoft version of "IBM Watson", where they renamed seemingly unrelated projects to Watson.

> It's not enough for the people who run a service to be good people – they also have to take steps to insulate themselves (and their successors) from the kind of drip-drip-drip rationalizations that turn a series of small ethical waivers into a cumulative avalanche of pure wickedness

I completely disagree with this guy. No amount of process or protocol can be a substitute for an actually decent organization culture; without the latter, everything falls apart no matter how good the former is.

Mostly just markdown-based skills. I've personally had more luck with harnesses, preconfigured permissions, and scripts to automate the frequent workflows, and the repo seems pretty light on that.

Building exactly this; in Mikoto Platforms, "Spaces" can be located on any physical node, and DMs are E2EE routed through multiple nodes

Slightly tangential, but this paragraph is the only one on the rules page with a "id" attr set, so that you can link to this specific rule

Then explicitly refuse service, instead of mocking your userbase.

Eh, AP has its own sets of problems (underspecified protocol, split-brained on discoverability, new developments are met with hostility in the community)

As a startup founder, your userbase is your god. Either treat them with utmost respect, or learn to explicitly fire your customers.

I don't care about the specific situation either way; What I am observant of is how the core team has handled their userbase and lack of protocol robustness.

and I actually don't hate that bit (I really like lexicons, although I might have approached it in a different way) - what I hate is the aggregation layer. I know that it is possible to have an AppView-less atproto app (e.g. RedDwarf), but I feel like much of the ecosystem still defaults to the assumption that it will go through the Bluesky AppView.

We already could've done that before, just throw a HTML file on a HTTP server on a cheap VPS

I don't even think having Jesse Singal on the platform is the problem (like it or not, I believe that all beings must have the right to communicate); the problem here was the communication failure when communicating this decision to the userbase. They could have just reiterated their rules and left it at that; instead, they chose to mock their userbase, write them off as harassment, and banned users left and right, abusing their position in network to censor people at every layer of the protocol.

We don't need "open" networks (Digital Panopticons), we need private networks.

How do you feel about the recent communication failures from the team to the userbase? As another builder of an open-source social platform, we must all understand that it is paramount for any company to not antagonize its customers, doubly so for a SOCIAL platform. I do understand that Bluesky and ATProto has to deal with a lot of baggage from both the old userbase and the new influx from the X/Twitter exodus, but engaging in user-antagonistic communication caused me to sour on the whole protocol.

Reminds of Unity features

I think the main advantage of the XML here is that the model is expected to have a matching end tag that is balanced, which reduces the likelihood of malformed outputs.

> After building Airbyte into a large open-source data infrastructure company...

Didn't Airbyte rugpull their license to ELv2?

This feels like an opposite of my experience. I find just having a bunch of Claude Code terminals running in the background to be the most "ambient", while I find autocomplete/auto-navigate the likes of Copilot/Cursor to be so annoying that I don't use any AI autocomplete now. Regardless, I think there's still a lot of room for structured AI programming flows, especially regarding semantic search, code flow tracing and intelligent find and replace.

Also (potentially) related?: https://en.wikipedia.org/wiki/Stargate_LLC

MMW: this will likely end up being the SVB for nonprofits

@dang likely a bot

> I think they’re recovering. They’ve learned a few lessons, including not to be too in hock to a few powerful and wealthy individuals.

I do not believe the EA movement to be recoverable; it is built on flawed foundations and its issues are inherent. The only way I see out of it is total dissolution; it cannot be reformed.

How long until people build EpsteinGPT that does semantic search on this?