227 Followers
224 Following
430 Posts

Web developer for projects for a better world.

This is a work account. Plenty of Drupal, PHP, web community and dev, and projects.
Personal account @ekes

Drupalhttps://drupal.org/u/ekes
Githubhttps://github.com/ekes
Gitlabhttps://gitlab.com/ekes

Status: 9M hits from AI crawlers, <0.1% human.
---

My first server was Apache on Suse Linux, in yr 2000. I've since deployed numerous webservers, many high traffic

Never guessed I'd one day be deploying a web project almost entirely for bots.

This is a view under the hood; the web is now more bot than human.

Our ideas, expressions, work, are being stolen for simulation & profit by a handful of companies, & at a scale comparable to deep-sea trawling.

It is our moral imperative to fight back.

@chgowiz

YYYY-MM-DD\THH:MM:SS±hh:mm[Area/Location]

It's as good a time as it gets.

RE: https://en.osm.town/@harry_wood/116562324497918231

So are the (Local Authority) Custodians of the address data able to license it? Shouldn't it be Open Data? I have to say I'm pleased The Netherlands has fallen on the correct side of this one (look at the addresses in OSM in NL).

@greg_harvey https://en.wikipedia.org/wiki/Subbotnik
I am told they are remembered very differently by different people.
Subbotnik - Wikipedia

I'm not sure people realise, but trying to review code rather than doing it myself requires much more cognitive power, especially when my trust of the job getting done properly is extremely low.

Yes one must start somewhere, but I've got *extremely* boring tasks of auditing certain engine APIs that are accessible to basically anyone and nobody wants to do it.
When this would actually be more impactful than following my "lead", especially if it comes out of nowhere and not ask me if I want help.

RE: https://mastodon.bsd.cafe/@stefano/116396440868847302

It's not like heavily upfront funded corporations haven't been known to be unreliable in the past. When the crunch comes capital wants to be paid.

And these particular corporations have already shown themselves not to be the most consistently ethical.

I'm sure there were reasons till a year ago we were pushing for things like more Open Source, and lower resource usage. What were they again?

@coldclimate Setting aside the important ethical and environmental considerations, what worries me is the dependency that is being created around these tools. My fear is that we will reach a point where only a very small number of companies can decide to raise prices or shut down services, and no one (or almost no one) will be able to do much of anything anymore. Once those skills are lost, it will be virtually impossible to recover them. Open Source models, although evolving, are not yet even close to the usability of the latest closed models, and there is little interest in their development (information I've been getting from developers). The conditions for a perfect storm.
We are aware of the outages yesterday and this morning, and have been working to find more resilient solutions. We have seen an increasing pattern of massively distributed scraping via residential IPs in ways that neither our Fastly CDN WAF/DDOS protection nor Human Security bot detection have been able to mitigate.

@liaizon Nope.

I see different actor and attributedTo for collaborative things: mobilizon, discourse. But attributedTo is singular.

Does an attempt at a search of the web for json strings. Ohh... could PeerTube potentially: https://github.com/Chocobozzz/PeerTube/issues/2704

ActivityPub Video `attributedTo` field malformed · Issue #2704 · Chocobozzz/PeerTube

What happened? When fetching videos, the attributedTo looks like this: "attributedTo": [ { "type": "Person", "id": "https://peertube.koehn.com/accounts/bkoehn" }, { "type": "Group", "id": "https://...

GitHub
@GeeBee I had to write a plugin for unknown wordpress sites. Making the output configurable by themes (many filters I recall) hurt my brain. Not sure many folk then actually configured the output (properly) afterward either.