I've added a few more bells and whistles to my agentic rube goldberg, but the gist is forgejo tag listeners triggering argo workflows to orchestrate
1. issue tag
2. write pr
3. testing
4. review+revise loop
5. merge mutex to ensure you don't get a merge storm
6. rebase and merge
I've been trying really hard to have it properly implement agentic identity where the pod gets a spiffe-attested token and then trades that for access to the vault secret for a project-scoped forgejo service account. I wish forgejo could configure a trusted external jwt signing authority so I could skip vault and the accounts.
The last piece has been using gvisor + kubernetes agent sandboxes. My fable adventure last week was having it debug the process of attesting and distributing workload identities for agents running in gvisor, as it creates a layer of indirection that confuses spire to the point it won't issue an ID.
A lot of people purchasing their products have a vague understanding of the problem they're trying to solve and an even worse grasp of how dbx solves it for them. I'm living this first hand.
They do hire plenty of American citizens, but the lengths they go to hire people on H-1Bs make me think they get something extra out of it. At FB you'd often see big boards of "public job postings" in internal lobbies that I can only assume were to comply with some arbitrary requirement.
Off the top of your head, could you describe a good/great ml ops infra setup? I've been reading what I can about state of the art ml ops after not being in the space for a few years and metaflow comes up a lot.
If I were interested in assembling an authoritative, up-to-date list of trusted CAs, would be reasonable to source lists from the major trust store providers and select only those CAs trusted by all of them? I know it's possible to be a lot more sophisticated and that even that can be flawed, but I'm hunting for a simple-to-follow criteria for now.
Clearly yes; my point was that the taste and decision makers that previously might've taken stronger cues from Twitter are now much less likely to see it as a venue for worthwhile discussions. I'm probably overvaluing that vs the harm of broadcasting puerile content, but I recognize this is fundamentally an unserious take.
My favorite pet theory is that Musk taking over Twitter is to disentangle the bureaucratic and journalistic class from the direct Tumblr/4chan content pipeline they've all been drinking from for years. The only way he could make it work was to make the space as distasteful as possible. I'm giving him way too much credit, of course, but I can dream.
I never saw this substantiated, but I heard that they had on-device systems that they used in emergencies for stuff that was particularly bad + easy to scan for. The time I heard of it being deployed was to prevent the spread of the Christchurch shooting video.
I don't believe this was in the original edition, but someone can correct me if I am wrong: he proposed the concept of "signifiers" to fill the definition that people were increasingly giving to "affordances".
A signifier helps to indicate the presence of an affordance that might not be immediately apparent.
I haven't read too closely, but no one is building meaningfully affordable housing in Atherton. Each unit in the development would easily sell for millions.
CA/the bay area should focus on packing high density housing closer to transit lines, not lifting apartments in areas that are an hour+ walk from transit to score political points.
This struggle reminds me of my personal battle with morning glories, nadinas, and bermuda grass, and a neighborhood effort to remove several ailanthus from along a local creekbed. Rhisomatic weeds are just a style of plant that we've never developed a good strategy for, outside continuous poisoning with something like glyphosate.
If you do ever see an ailanthus seedling, dig it out if you can. They're terrible and I've even seen them crowding out young redwoods here in northern CA.
https://extension.psu.edu/tree-of-heaven
True, though his videos often fit the criteria of including "tactile" or "natural" sounds and he seems careful not to add anything that would get in the way.
The article lacks a date, but I assume the best season for nuts is in the late fall, October through November. Does anyone know otherwise? I was planning on going foraging for nuts all last year but couldn't seem to time it right.
Would you be open to sharing some details about your project(s)? Things like costs, who you worked with, guiding principles, etc. would be amazing. I feel like maybe finding the kind of people that will do high performance work is just outside of my budget, so I'm curious what you did and with what resources.
It seems like a series of guiding concepts and examples for builders or clients. Custom home building seems to be the main venue for this at the moment, but a layperson like myself can use this to try and recognize a good house or think about how I might retrofit my own. It's all too expensive at the moment where I live, but I can dream and save my pennies.
I might get a better answer on GBA, but any building performance folks you'd recommend connecting with in the SF Bay Area? I'm not in the building industry at all, but love hearing talks and daydreaming about changing careers. Finding some material for folks outside of the PNW, New England, and the South West would be amazing. Temperate climates like the bay seem especially challenging to get specific info for, so maybe it's a case where I need to understand the principles well and apply it myself.
Love the book by the way; it's what I wish I'd had when I was looking to buy my first home. I'm going to be gifting it to a few people in my family over the next few years.
In some cases reintroduction can be positive because certain species were selectively hunted, the factors involved in habitat loss have been reversed, or the species might play an important role in habitat restoration.
I think this is true, but the balance has shifted dramatically and will likely continue to do so indefinitely. It used to be that the broad range of opinions and personalities balanced out /pol/ when you were visiting other boards, but it really is getting to the point where every board I visit is skewed by at least some amount of antisemitism, conspiratorial thinking, christian fundamentalism, or pro-nazi thinking. The most amusing thing to happen lately on /pol/ has been their struggle to cohere a consensus opinion around the war in Ukraine. It might be the first broad division I've seen there in a while.
I'll add my own annecdote to this article. We had a wasp nest under the eaves of the house and I managed to successfully move them into a little hutch I built near the garden. I can't say I directly observed them saving the tomatoes, but we did have fewer tomato hornworms than in previous years.
They were also really docile and tolerated a good deal of me fumbling around with the nest. I waited until it was cold out and covered up pretty thouroghly, but they were mostly content to let me flop them around with a set of long BBQ tongs.
1. issue tag
2. write pr
3. testing
4. review+revise loop
5. merge mutex to ensure you don't get a merge storm
6. rebase and merge
I've been trying really hard to have it properly implement agentic identity where the pod gets a spiffe-attested token and then trades that for access to the vault secret for a project-scoped forgejo service account. I wish forgejo could configure a trusted external jwt signing authority so I could skip vault and the accounts.
Here's the inspiration for the auth model I've been trying to implement: https://datatracker.ietf.org/doc/draft-klrc-aiagent-auth/
The last piece has been using gvisor + kubernetes agent sandboxes. My fable adventure last week was having it debug the process of attesting and distributing workload identities for agents running in gvisor, as it creates a layer of indirection that confuses spire to the point it won't issue an ID.