terry matula

43 Followers
104 Following
216 Posts

Aspiring Hallmark Christmas movie soundtrack composer.

* Pro PHP developer since 2009
* Started the #laravel Austin meetup in 2013
* Wrote the "Laravel Application Development Cookbook" in 2014.
* Worked at various startups
* Currently software engineer at a healthcare company.

:verified:
Blog?https://terrymatula.com
Hub of Gitshttps://github.com/matula
BlueSkyhttps://bsky.app/profile/terrymatula.com

RE: https://phpc.social/@Crell/116262239935723893

We bought a used Hyundai Kona EV, after having a Leaf EV for many years. Easily goes 200 miles on a charge, and with planning, much further if you find 1 hr fast chargers on your route. No subscriptions, and I think our monthly electric bill has MAYBE gone up $10ish on average.

Still have a gas minivan, but rarely gets used these days. The only downside is the $200 a year we have to pay Texas, to offset gas tax.

little thing I was curious about: https://matula.github.io/furry.html
Epic Fury

99% of my usage is "turn these files into 320k mp3s" or "convert this file to an mp4"... And that's solid. But I also noticed it'll attempt other ffmpeg flags as well, like being able to edit the length of a file... Or convert to gif. I added some volume level shots as well, but haven't really tested it.

I'm not really sure what ffmpeg the Qwen model has by default. But might be cool to custom train one and get the model size as lean as possible.

https://github.com/matula/ai-ncoder

I hacked this together months ago because I found myself converting a bunch of old audio files, and I was continually googling #ffmpeg flags. I thought it would be cool to use natural language for transcoding, but wanted it to keep it local, and workable offline.

So I grabbed the smallest, usable model and hacked together some Python and prompts... And it just kind of worked.

I thought I'd clean it up and share.

GitHub - matula/ai-ncoder: AI-ncoder

AI-ncoder. Contribute to matula/ai-ncoder development by creating an account on GitHub.

GitHub

The guy on the left was a US Marine Lt. Colonel who was arrested and convicted for selling missiles to Iran.

The guy on the right is a Fox "news" *snicker* "military analyst" who says Iran shouldn't have missiles.

They're the same guy.

Ha! Just yesterday was musing about how I hadn't seen any Anthropic traffic. At least their UA is easily blockable.

@bobmagicii I could just be lucky, and looking at logs I don't see any "Claude" ua. I do see "amazonbot" and "GPTBot" among others. If Anthropic is hitting my site, they're not slamming it at least.

Looking at search referrers, it's interesting I don't see Google. Just DDG and Bing. Not sure if that's just because of the type of people who might search for obscure Houston bands, or Gemini screwing up search.

@bobmagicii it's a bit janky, but I pass the IP to the Team Cymru service to get the ASN. And I'm storing the asn to a blocklist.json file.

My biggest issue was with Tencent Chinese servers, with multiple IPs but they all had the same ASN. This way I can block ALL of them with a single value instead of hundreds/thousands of different IPs.

For AI bots that have "good" user agents, I have a separate system that parses the ua and blocks based on that. Most big AI companies are good about it.

I learned about Team Cymru, and their IP services. Pretty cool, and "free". Using PHP's dns_get_record, I'm able to make a super quick query for an ASN record, and block the ones I don't want.

I'm no security expert, but this seems to be working well.

This whole project is just pleasant to work on... no frameworks or database or external dependencies. Just pure PHP, with a smattering of CSS and JS.

Kind of wild to see all the Mastodon/Fediverse instances that immediately ping my site after posting a link or even replying to a post with a link.

And then right after that, all the "exploit" pings.