The world uses 102 million barrels per day.

https://www.worldometers.info/oil/

And the article says they’ll be releasing it over the course of 120 days. Maybe it’ll bring the price down by a few percent, not enough to be noticeable.

It’s just not practical to store enough oil to make any long term difference, we use too much of it and it’s too valuable to leave lying around unused.

20 hours of work turns into 20 minutes

The gains, where they exist, are nowhere near that much. In some cases, it makes developers slower (even though they think they’re a bit faster):

we find that when developers use AI tools, they take 19% longer than without - AI makes them slower.

https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/

PieFed.social used to have 4 GB of ram so it’s totally doable, with a bit of tuning. I will tweak the default compose file to have smaller values because all servers are small at the start.

Yeah that sounds better.

I normally use https://pgtune.leopard.in.ua/. Based on that I recommend:

shared_buffers = 1GB
effective_cache_size = 2GB
maintenance_work_mem = 256MB
checkpoint_completion_target = 0.9
wal_buffers = 16MB
default_statistics_target = 100
random_page_cost = 1.1
effective_io_concurrency = 200
work_mem = 16MB
huge_pages = off
min_wal_size = 1GB
max_wal_size = 3GB

That’s weird.

How much RAM does your server have? Maybe it’s running out - postgresql tends to start out using hardly any RAM and then it grows over time until it reaches the limit set by it’s configuration. Check what that is - it should be no more than half your RAM. In the compose.yaml near the start there is the bits you can tweak for postgresql.

My partner has been looking for a job for a year.

Two weeks ago she supplied her passport to the main job seeking website in our country, which gave her profile a blue “verified” tick. Since then she’s had 4 job interviews, more than she’s had in the last 6 months.

Recruiters are being flooded with low quality applications, many of them AI-generated bullshit, many of them from overseas. Proving you are who you say you are and that you are in the right country seems to be necessary, now.

!general@activitypub.space, that one. Yep.

There is a convention of putting a FEDERATION.md file in the root of your project (there is a FEP about it but I can’t remember what it is, heh) which lists the FEPs your project implements.

e.g. https://codeberg.org/rimu/pyfedi/src/branch/main/FEDERATION.md

There is no standardised API endpoint for discovering this info so basically you just need to lurk in all the dev spaces you can and watch what they’re talking about, stalk each other’s codeberg/github issue queues and PRs and just cope with the jank.

Yes theoretically you could have any type in there but votes are the biggest inefficiency in the system so I’ve only implemented it for them.

There are some nodebb forums that federate with us, which are intended for AP dev discussions - !technical-discussion@activitypub.space for example. Not very active. Also some Matrix channels, again not exactly humming. Everyone is pretty focused on their own projects, really.

Maybe.

I can’t find any gavinmooney profiles on any socials… even x dot com.

It’s probably just AI generated bs.

Generally, solar takes 10+ years to break even in a residential situation, I can’t see how things would be 10x cheaper at the TWh scale.

Go to https://piefed.social/user/settings/filters and check your Instance filters. You’ve blocked lemmy.ca so none of your posts in lemmy.ca communities get sent there.

I will see what we can do at the UI level to warn people about this.

There is, yep. You can just walk away, but that leaves everyone thinking that you learned nothing and still hold the silly views you expressed earlier.

There is a discussion but it’s an email chain between myself, Lemmy, Mbin, NodeBB so not really out in the open.

I’m thinking of using the normal OrderedCollection structure that AP has - https://www.w3.org/TR/activitystreams-vocabulary/#dfn-orderedcollection. An object of type ‘OrderedCollection’ with an array of ‘items’.

I have never seen anyone use this!

If you have 100+ notifications you probably are getting dogpiled for saying something controversial. Rather than replying individually to each flame you can just use that function which puts a message at the top of the post and locks comments.

It’s a de-escalation tool.

You’re right, to be more similar to the ActivityPub style of doing things the object should be a “type” of “Collection”, not just suddenly be an array instead of an object. I will migrate PieFed over to that instead.

Yes, well, that’s what this feature is for:

That sounds pretty fun, I like your extremist take on things.

If you want to federate votes in an efficient way, consider doing them in batches, like this - https://codeberg.org/rimu/pyfedi/src/branch/main/docs/fep-4248.md. Sending batches means building up a queue, etc which is a pain in the ass so you might want to just receive batches and send individually.

You can do nested comments more efficiently by doing a single query for all the comments on a post and then constructing the comment tree in Go, rather than using multiple SQL queries. It’s a bit tricky to code but you can copy PieFed’s code - https://codeberg.org/rimu/pyfedi/src/commit/68d6799be5c321f9abf84d136cafef6de6e3aafb/app/post/util.py#L141

Man, I am a cranky bastard sometimes. Shouldn’t post when I’m tired.

How about - the origin server, the place where the post was originally created, gets all the metadata (og:image, og:title, etc) and includes that in the Activity that tells other instances about the post? That way there is only one request made to the link and receiving instances can use that to make their preview?

It does mean that receiving instances will need to trust the sender so there’s potential for some misrepresentation but that seems acceptable to me.