<unnamed>

Public Key

npub1nsp3hg75ge84dn5fttlkpxpxnj0dfl697239dtyz4js3mdw2jllqf8ahke

Profile Code

nprofile1qqsfcqcm502yvn6ke6y44lmqnqnfe8k5lazl9gjk4jp2eggakh9f0lsprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntqyqq2h2347

Publishing to

relay.momostr.pink

Last Notes

2026-08-18 16:20:28 CEST

- reply

I think "trusted platform" and "secure the supply chain" open the door for foundations to receive funding. I am using the "trust platform" but I can't tell what it's solving for me. (I don't need anyone to explain it to me.)

2026-08-18 15:38:20 CEST

- reply

24 hours later and #GitHub's status page lies about how long these services were down.
0 hours to 2 hours and some change across the board.
Mr Shu's page shows >7 hours across the board, which is what I saw, since this spanned 90% of my workday yesterday.

2026-08-17 20:36:13 CEST

🟢 Keep https://www.githubstatus.com close.
🔴 Keep https://mrshu.github.io/github-statuses/ closer.
#GitHub #GitHubStatus

2026-08-14 16:07:49 CEST

RE: https://mastodon.social/@webology/117091821643147401
⏰ Office Hours today in about ~5 hours.
Starting 30 minutes early, I'd love to talk with anyone interested in the non-boring parts of packaging.
I have several projects to package, and I know things have changed in 2026.
I'm also dusting off my python.org commit bit to try and clear some random issues that are junking up the pipes. https://github.com/python/pythondotorg/issues?q=is%3Apr+is%3Aopen+author%3Ajefftriplett
#note10sh…3nat

2026-08-14 05:50:07 CEST

🎙️ I'm hosting Office Hours tomorrow, Friday, August 14th, 2026, from 2pm to 4pm CT aka https://time.is/0200PM_14_August_2026_in_CT?Jeff%27s_Office_Hours
DM me for details or use the gist we always use.

2026-08-06 17:02:35 CEST

RE: https://mastodon.social/@djangonews/117049132052049599
It's official: @npub1l0c…kt92 is accepting applications to hire an Executive Director.
If you want to join the DSF's weekly office hours to discuss, we hold them every Wednesday at these times: https://tools.jacobian.org/dsf-office-hours/
#note1m3z…s9cu

2026-07-31 14:45:51 CEST

Now that GitHub's new UI has had several years to settle in, my favorite feature is how pressing F5 on the pull requests screen magically shows open pull requests, whereas clicking into the page, >90% of the time, shows fewer results.
I ****love**** that feature. I see it multiple times, every single day.

2026-07-31 14:43:33 CEST

- reply

for me, I deeply love the iPad Mini over the other formats. I have owned three of them vs. one mega 13" when they first came out. Why does Apple not use the new hardware and refresh them all at same time, vs the Mine being three years+ hardware and only updated every 4 or 5 years it seems.

2026-07-29 21:29:44 CEST

- reply

Hipster!

2026-07-28 13:27:02 CEST

- reply

I'm sorry. The no-LLM crowd calls me AI-pilled and yet the gap of pro-LLM usage to the real AI-pilled/AGI is already here believers “gap” is still a mile wide. It's frustrating but I am still going to social media and discord how I want to.

2026-07-25 16:38:55 CEST

Today is a great day to check your database backups.
Mine are doing their jobs, except for the PostgreSQL database I launched six months ago, which I used to access pgvector.
I forgot to set up backups, and things are fine, but it was only a matter of time until I needed to pull a backup.
One advantage of Coolify is that they automate backups and even support any S3-storage api like R2 storage. They even make it easy to restore from a backup to a different server.

2026-07-21 14:05:27 CEST

- reply

I'm all for a good study, but it reads like taking @npub106g…uas2's https://whereabouts.earth/daily/ app, adding an AI mode where you ask people to use the tool, but you give them wrong answers on purpose for your study when you know the tool will mostly give the wrong answer.
I agree on narrative, but I fail to see "research" 😬

2026-07-21 13:14:21 CEST

- reply

Seems like a bad study if only 44% can get it correct with or without LLMs. Especially if you tell them to use a bad model you know will fail on subjects they know nothing about. Paying them incentivized them to care slightly more but clearly not enough to abandon the tool you asked them to use. 😬

2026-07-20 21:16:39 CEST

- reply

Definitely printing this once I get my printers recalibrated.
I built my own Franken Steam Deck and needed a solution to get around its 128 GB internal drive, and this might be perfect for it.

2026-07-20 14:45:03 CEST

- reply

I really struggle when people pretend that Sweeney was ever aligned with gamers vs. adding to Epic's bottom line and another billion for him to launder. (Not that you were pretending or anything like that.)

2026-07-16 15:01:50 CEST

- reply

took a minute to find it, but 37% + a few more increases. I noticed a third bump on my bill without any announcement. I'll have to tear into their API to figure it out, but too many other things to do before that point.
This was mostly targeting their US servers, btw. I know a few people who shifted to European servers, but several shifted back again from whatever recent legislation made them bail again.
https://www.hetzner.com/pressroom/statement-price-adjustment/

2026-07-16 14:57:48 CEST

- reply

Re: Hetzner - I would understand an increase for new servers, as they did for their second or third price increase this year.
It was pretty rich seeing them continue to post screenshots of $5 servers on socials that no longer existed for months on end. I think they are still up.
I understand new servers cost more money, but they already baked in the price, and it's greedy AF to increase prices this much.
https://files.mastodon.social/media_attachments/files/116/929/759/588/310/946/original/d102c43786f08328.png

2026-07-16 02:51:48 CEST

I really liked Hetzner, but they seem to be taking too many plays from DigitalOcean's playbook. Between >2x'ing their pricing over the last year and now discontinuing their referral program without even the decency of an email to let me know.
What are the non-Hetzner and non-DigitalOcean users out there who need 3 to 5 servers?

2026-07-12 23:11:37 CEST

- reply

Sounds cool. I will definitely check it out after you release it.
I have two or three services I use to sync many calendars back to one account, and I think I might be better off getting rid of them and throwing some tokens at it.

2026-07-08 22:47:24 CEST

RE: https://mastodon.social/@djangonews/116886123253151601
Last call to give #django feedback in this year's Developers Survey.
I, too, need to fill it out! 😬
#note13xx…qhnt

2026-06-30 15:06:37 CEST

- reply

I just can't type on those, but I'd love to see a write-up if you do.
My everyday carry keyboard is Keys to Go 2, which is amazing, but it doesn't magnetize to the device. It's great for planes, coffee shops, and conferences.
https://www.logitech.com/en-us/shop/p/keys-to-go2-universal.920-012920

2026-06-30 14:56:37 CEST

- reply

That gives me an awful idea with my 3d printer + magic keyboard + magsafe or magnets... 🤔

2026-06-26 20:51:22 CEST

- reply

I'm team warn loudly vs. not being able to log in to your website without a few non-obvious setting changes.

2026-06-26 20:50:39 CEST

- reply

To pick on Django: the average person deploying their first Django website should not encounter any tripping hazards that would prevent them from seeing a working website.
If they are going to accept credit card/financial details, then yes, but for a blog... nope. We uniquely make this harder than most frameworks do, and we are not better for it.

2026-06-26 20:11:42 CEST

- reply

Instant deploys have bubbled up on my list of problems I want to see solved.
We have solutions for web hosting, and the rest is solved with better defaults and not confusing "does not work by default" with "that's just good security" when we have overplayed what good security is and means.

2026-06-26 19:26:20 CEST

- reply

I don't think we skipped it. There are many web UIs for this stuff, and the best ones were originally chat interfaces.
The letdown for me is that they have to be hosted somewhere, and they have to have reasonable access, which isn't always feasible.
For you and me to host our own OpenCode server is fine, but for the average person, they don't have the skills to do it. Tailscale is wonderful, but a really hard topic to cover with someone who is new to everything.

2026-06-26 15:10:44 CEST

- reply

I agree that it has nothing to do with retro computing. TUI tooling has gotten really good, and no one has figured out a GUI pattern that works with how many of us drive LLMs.
Eventually, a pattern will emerge, and it will be obvious what that looks like, but everything either resembled an IDE or an Email app, and we don't want or need that.
There's a huge architectural difference, too. Many of us are running a TUI remotely, so it stays persistent, and we can be mobile.

2026-06-19 07:15:16 CEST

- reply

I have never worked with a client who used t-shirt sizes but it makes a ton more sense than story points.

2026-06-17 15:36:00 CEST

RE: https://mastodon.social/@djangonews/116765592686029364
Thanks to the support of six of our sponsors, the Django Software Foundation has secured initial funding to support an Executive Director.
This is the very beginning, and we have a lot of planning and discussions to go, but this is a big step towards hiring an Executive Director.
#Django #Python #DSF
#note13ha…xdmj

2026-05-30 02:07:34 CEST

- reply

4.8 is a bummer so far. I have some fun with it, merging PRs it wasn't supposed to, and other "fun" issues for a Friday.

2026-05-01 16:13:17 CEST

- reply

Blocker for me has been not having a good Actions solution. I know it may be hard to do at scale, but I'm more likely to pay and use it if they can figure it out. I tried to run my own runner ~5 months ago and it was sadness.
https://docs.codeberg.org/ci/actions/

2026-05-01 15:39:01 CEST

RE: https://chaos.social/@carlton/116497687371540613
🤔 I have seen this topic come up a lot this week, and what I find interesting is that we somehow gloss over https://codeberg.org (a community/public GitHub alternative) and https://forgejo.org (self-hosted and powered by Codeberg).
Maybe "community hosting" could start with Forgejo, but Codeberg is good enough to get off of GitHub today.
It's not a good fit for large projects yet, but I think we should talk more about what they need to get there.
#note1ggt…n5jh

2026-05-01 06:14:47 CEST

💧 Nothing like hearing water pouring into your basement (luckily unfinished) on a Thursday night, two hours before the hardware store closes, to get the blood flowing.
Two trips to Menards and $40 later, it's fixed better than it was until I can get a plumber out to give it a more permanent fix.

2026-04-30 20:11:59 CEST

DjangoChat Podcast: E201: https://micro.webology.dev/2026/04/30/djangochat-podcast-e/
This week, I joined the DjangoChat podcast episode discussing DjangoCon Europe, recent developments in Django, AI workflows, and shared recommendations for projects and books.

2026-04-29 15:50:56 CEST

- reply

I had one that was a food blogger in Brooklin, who, by the end, was in space fighting aliens. While that sounds kind of cool, that was not the plot, and context drift + hallucinations caused us to go off-script.

2026-04-29 15:49:17 CEST

- reply

Excellent advice. I used it's vision mode on a bunch of photos and with pydantic-ai (my hammer), I extracted a bunch of data. You are right about it being slow.
I like deepseek-v4 too. It has been really impressive with creative writing. I have a weird experiment I have written to test how well this tech works, that effectively writes a five to x chapter book just to see the character/plot drift, and DeepSeek has always shone for this.

2026-04-29 14:13:47 CEST

- reply

Circling back to this. What a brilliant headline.
Semi-related: I moved some of my daily agents to Ollama-hosted models as a baby step toward testing local models. I was using GPT-5 (mini|nano) and so far so good.

2026-04-24 14:12:19 CEST

🚀 We shipped Issue #334 of @npub1qsd…3jep on @npub1hhf…qrnt this week for the first time.
📊 If you subscribe, did you receive it?
https://django-news.com/archive/issue-334-new-look-new-home-same-everything-else/
#Django #DjangoNews #Python

2026-04-22 16:35:24 CEST

RE: https://techhub.social/@pycharm/116448755917975441
➡️ Please help us spread the word, and this is an excellent chance for you to try out PyCharm and help #Django out.
#note1ptp…vqvg

2026-03-28 18:21:52 CET

- reply

My cheat code for the calendar was mostly read-only, but I turned it into markdown bulleted lists by day. LLMs a good at reading it and it's easier thean trying to parse json or CSV which eats a lot of tokens.

2026-03-28 17:59:36 CET

- reply

Super cool.
I have been really enjoying building these lately. My most useful helps me with time logging and reading my calendar. It seemed really dumb until I gave it a tool to read/update it's own memory files and track past actions/corrections. Now, it's strangely a joy to work with.

2026-03-27 16:21:25 CET

- reply

I think the whole federation means something different to everyone and how that works is what make it feel weird.
As long as federation means, we can stand up some hubs for discovery, stars, etc then we can land in a good place. If that data is federated than great.
But federated as liked or bookmarked via a social app-like experience is as good as never happened.

2026-03-27 14:33:10 CET

RE: https://techhub.social/@ironicbadger/116301246624014978
I feel this, but I also fear 100s, 1000s, and 10ks of self-hosted servers with different rules and nuances just to figure out how to post an issue or fork a project, and I fear we are going to leave new contributors even further behind.
I struggle to say, let's all hop to https://codeberg.org or another platform, since that probably kicks the can down the road until it's too big and fights the same fight.
#note1c43…y7cv

2026-03-27 14:04:04 CET

- reply

Looks neat. Does Go have something like Pydantic AI? Between it's OpenAI and Anthropic API support, it really makes quick work of weird thid party API providers which I seem to like using.

2026-03-22 21:22:43 CET

- reply

Super cool. I haven't tried to voice clone yet.
I never landed on a useful MCP implementation. I still think it'd be useful to run on a production website, so I could connect to it, but I never got it working reliably.

2026-03-21 22:51:18 CET

- reply

I upgraded my small fleet of Macs and it was very uneventful.
I was expecting my screens to melt and my hardware to stop working and other than things being a little more round, I struggle with why this is a hill, but there were a few quirks which have all be resolved.

2026-03-18 15:04:04 CET

- reply

I don't think it makes you anti-kid to not go, but if you like your friends and want an excuse to 🙄 with a group who probably feels the same, you have a free social engagement with no pressure to actually gift anything.
We get to opt out of these because we live 200 and 600 miles away from both sides of the family.

2026-03-18 15:02:11 CET

- reply

So, person with two kids, and we have opted them out, and it was pretty eye-rolling. I think people are excited about having a kid, and it's an excuse to celebrate. We only had one baby shower, which is also not the norm these days. Most people have them per kid and we were like why?
So... Friends meet for lots of dumb reasons like friendsgiving or to celebrate a break-up or divorce or engagement or because it's Friday night or they have a babysitter and lots of random things.

2026-03-11 15:25:39 CET

- reply

oh sorry. I completely forgot how bad that gets 45 secs in or so. Sorry about that 🙈

2026-03-11 15:24:14 CET

- reply

https://www.youtube.com/watch?v=hJ9yBgTp9UQ came to mind

2026-03-11 15:08:29 CET

- reply

My guess is 13,00

2026-03-11 13:18:54 CET

🎉 Happpy VeldaTV day to those who celebrate it.
https://djangotv.com/videos/djangochat/2026/from-kenya-to-london-velda-kiara/
🎉@npub1m3d…w9kh
#Django #DjangoChat #DjangoTV #Python

2026-03-05 18:52:59 CET

- reply

I'd love to pinch and zoom on my laptop's screen. Same for moving windows around or interacting with my iPhone via the Simulator or iPhone Mirroring (mostly used when my phone is somewhere but not on my desk or pocket)
I am mostly poking fun at you both (podcast) being "wrong" because, grim or not, it's a useful feature that other hardware manufacturers have mastered. It's also one you could ignore.

2026-03-05 17:50:27 CET

- reply

They touched on being anti-touch screen on today's Abstractions podcast. They are wrong, btw, about it, but I'm sure shipping would have a touchscreen, which means one could turn it off.
https://mastodon.social/@npub1rcp…82r4/116175791381993128
cc @npub1hh3…5gzz and @npub1rcp…82r4

2026-03-05 15:46:11 CET

- reply

Looks like cpu cores, gpu cores, and ram + an A-series chip vs. the M-series which which I didn't realize they still made A-series chips.
https://www.apple.com/macbook-neo/compare/
I have said this for a decade+, but they really should ship with touch screens too.

2026-03-05 15:26:28 CET

- reply

battery + screen + keyboard + touchpad for a one- or two-off machine seems fine to me.
For you and me, we probably have a spare 8k KVM switch just sitting on our desks because it seemed like a good deal... and now I just outed myself, but I almost bought a Mac Mini because they were $300 off before #ClawSanity lit fire.

2026-03-05 14:56:11 CET

- reply

I suspect the OpenClaw people are going to buy them up, given their cost, screen, and power vs. a Mac Mini.

2026-03-04 00:08:42 CET

- reply

Touch displays. Why can't we just skip ahead?

2026-03-01 19:04:50 CET

- reply

A concussion does not sound fun. I hope you are alright.

2026-02-26 16:47:04 CET

- reply

this feels like how our generation sets up playdates for our LLMs. 😂

2026-02-26 16:12:05 CET

- reply

Speaking of model hype (me, not you), I am excited but a bit time-poor to try out https://ollama.com/library/qwen3.5 and https://ollama.com/library/lfm2, which I'm seeing deliver better-than-I-thought-possible results for local models. Not sure if you are still dabbling locally, but I hope to have time and RAM to try them out this weekend.

2026-02-26 15:42:53 CET

- reply

It's all good. There are three parts to it. They keep filing as they figure it out. Overall, I am impressed by it, but I haven't read part three yet.
At the end of the day, these are just tools that people can do good and bad, but mostly in the middle. And it sucks to see them used for bad and causing grief.

2026-02-26 15:07:09 CET

- reply

Folks, just a reminder not to take legal advice from people from the internet who are not lawyers. They very highly highlighted and clipped three summaries from three cases which have no bearing on the recommendation.
Check out https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf and you'll see that the claim is largely 💩
They recommend not allowing 100% generated project copyright, but our existing laws allow everything else per their guidance/recommendation. Read the PDF.

2026-02-22 15:10:58 CET

- reply

Maybe. I think that's only a phase, though. Most of what the second model checks, the one-shot models can already do before they give you the result.
I think model collapse and model draft are different things and are largely avoidable. I admit I don't read whitepapers much, but it's well documented, and there are steps to avoid the issues. So I'm not sure it's inevitable?

2026-02-22 14:55:28 CET

- reply

Ollama is good at tool calling which honestly the biggest inflection point (IMO) in this tech. So having a model that can do things that you give it tools for changes everything.

2026-02-22 14:40:36 CET

- reply

This seems similir to https://www.cerebras.ai business model. I believe they build their own hardware aimed at running LLMs extrenely fast. Not quite the same as burning AI to a chip, which I am really fascinated by.
@npub122k…wx8u has a few demos of using it and I think it's 2k to 3k or more tokens per second.
By the end of 2026, I bet most, if not al lof the frontier models will run like Cerebras does.

2026-01-25 13:27:00 CET

- reply

Totally not the same thing, but when dependabot alerts ping me for some random JS vulnerability that has no impact on a static website gives me similar feels, and it's all annoying and unnecessary AF.

2026-01-22 20:58:06 CET

- reply

After my P1S came in a few weeks ago and I did my first day of prints, I checked and their sale was $300 off. I nearly ordered another one because it's so much faster to print now and so easy to keep both running for hours at a time.

2026-01-22 20:52:02 CET

- reply

Now that I own two 3d printers, I see the appeal of three or more. 🤔

2026-01-16 19:53:36 CET

- reply

that looks wild. Touch screen keyboards scare me, but it's a neat setup (and I'd be using an external keyboard)
Why hasn't Apple done this yet?

2025-12-23 16:20:03 CET

- reply

oh wow. Pictures at some point I hope.

2025-11-23 07:11:09 CET

- reply

I review everything via Git Tower or via PR.
This is a case where it's a bad default IMO, but I still should have caught it. https://django-q2.readthedocs.io/en/master/configure.html#max-attempts

2025-11-23 06:04:25 CET

🤖 How I Accidentally Spent Over 62 Million OpenAI Tokens
https://micro.webology.dev/2025/11/22/how-i-accidentally-spent-over/

2025-11-16 18:36:40 CET

Mozilla builds products for Google. Not to serve you. This hasn’t changed in decades despite some really amazing and caring humans working in the trenches there.
The companies survival depends on it until they figure out how to be profitable with something that’s not Google. But they haven’t and there are no signs that they ever will.
What Lobster commenters want: Adblock… https://mastodon.social/@lobsters/115560442394127527

2025-11-14 16:28:40 CET

- reply

I'm not a fan of the 3d Metroids, either. That said, it looked fun, but I know the controls will annoy me.

2025-11-06 16:06:57 CET

📺 DjangoCon US 2025 videos are up 🎉
This also caps off a week's worth of updates on DjangoTV, including the dreaded, very random 153 errors, which should finally be gone.
https://djangotv.com/videos/djangocon-us/2025/
#DjangoConUS #DjangoCon #Django #Python

2025-10-29 16:02:16 CET

- reply

When the PSF closing down their grants program, the DSF has start receiving more grant requests to help pick up the slack. We can use your help!
Big picture: The DSF operates on about 1/16 of the annual budget of the PSF, and we haven't made our goal for the year.
If your company uses Django or if your AI company uses Django to tout how good your AI models are, please support the Django Software Foundation. 🙏

2025-10-29 15:55:58 CET

👋 Friends, #Django's fundraiser is running through November 11th (kindly ignore the 9th, that's being fixed) and the foundation could use your support to continue our mission.
While I happily support the PSF in their mission, this is the DSF's biggest fundraiser every year, and even if you aren't a JetBrains user, our foundation could use your support.
Support the DSF via PyCharm:
https://www.djangoproject.com/weblog/2025/oct/23/pycharm-django-annual-fundraiser/
Support the DSF directly: https://www.djangoproject.com/fundraising/
#Python

2025-10-27 14:27:46 CET

RE: https://mastodon.social/@djangonews/115446256829600290
🎙️💬 And I'm on @npub120g…wm7n talking with @npub1mjx…httg and @npub1mfs…ngvj about the latest Django Survey trends with a WHOLE lot of side quests.
I hope you enjoy it!
https://djangochat.com/episodes/django-survey-2025-jeff-triplett
#Django #Python #DjangoChat
#note1xt5…32cf

2025-10-25 19:58:16 CEST

- reply

When the LLM has the ability to read the solution its a hard game to win.

2025-10-22 22:02:02 CEST

🤔 If you’re thinking about running for the Django Software Foundation Board of Directors (aka joining me!) but aren’t sure, please reach out to me or any of the other directors. We’d be happy to chat.
There are 3️⃣ three seats open this year, and I’d love to see a more diverse range of candidates step forward.
🔗 Call for Nominations: https://www.djangoproject.com/weblog/2025/oct/11/2026-dsf-board-nominations/

2025-10-20 20:20:58 CEST

🤖 I was lucky enough to be involved with the beta of Claude Code on the Web (works on your phone) and it's one of my favorite ways to use Claude.
You can have it work on a feature and open a pull request and resume from your terminal if you want to. Or not because most features I have been working on with it, doesn't need it.
https://www.anthropic.com/news/claude-code-on-the-web
#ClaudeCode

2025-10-19 14:58:04 CEST

- reply

I wonder if your CalDigit is bad? I had the opposite luck with other docks until I landed on my TS3. I tried out 4 or 5.

2025-10-16 18:57:49 CEST

🎉 I hadn't kept up with what Django's Steering Council had been up to outside of random board updates, and I'm happy to share how impressed I am with how transparent they are.
They are meeting several times a month and sharing their minutes as they go. https://github.com/django/steering-council
They are even sharing on the Django Forum to bring more visibility to what they are doing https://forum.djangoproject.com/t/django-steering-council-meetings-2025/38306
These are very welcoming changes. 👏👏
#Django #Python

2025-10-14 02:32:01 CEST

- reply

This is good https://tedium.co/2025/10/13/omarchy-linux-distro-commentary/
> Tech should be a respite from stuff like this. Instead, software has to be considered in terms of its ethical lines. For me, no matter its technical benefits, the “opinionated” Omarchy fails the test.
> There are plenty of other distros out there that don’t make me feel like I’m co-signing onto an ideological perspective that I otherwise would go out of my way to avoid.

2025-10-14 02:30:19 CEST

- reply

For anyone wanting a fork sans Hamburger Helper aka asshat, ArchRiot appears to be that. https://archriot.org
I couldn't get it to boot into a GUI when I tried it a month ago. I might give it another shot, but I'm probably going with a distro that's a glorified SteamOS desktop because I still don't think the Linux on a desktop is all there yet.
I have tried for over 30 years at this point, and I still haven't found what I'm looking for despite running it for years 😬

2025-10-14 01:55:06 CEST

Times I have been asked if I have tried Omarchy today? 3
Times I have tried it: 1
Times I have installed a Linux Desktop only to discover that Linux might work on a desktop, but it still sucks: every time and going back to the late 90s.
No hate, but just because someone discovers a tiling windows manager is cool, but it's hardly something to write home about.

2025-10-10 17:53:12 CEST

🤔 What's with all the framework (non-laptop) hate lately?
🤷 Just because React is overkill for most projects, doesn't mean all frameworks are bad.
If your default reaction to someone using a framework is that the underline technology isn't that hard, and they should just suck it up, then you have exceeded your gatekeeping quota for the year. Which should be zero, btw. 0️⃣ That's spelled ZERO.

2025-10-10 16:37:51 CEST

It's SOOOO... CLOSE....
The Django DEP 💯 (PR number) is so close to be accepted.
The goal of this DEP is to shorten Django's default cli command from `django-admin` to `django`.
https://github.com/django/deps/pull/100
Back story: https://micro.webology.dev/2025/01/08/rethinking-djangos-commandline-tool-why/
#Django

2025-10-07 01:57:25 CEST

RE: https://mastodon.social/@bagder/115241241075258997
It's refreshing to see someone knowledgeable use AI assisted tools to find more than 70 bug fixes for the Curl project.
Daniel has been transparent about his struggles managing the project because of bad AI usage, so this isn't coming from someone who hasn't seen it all.
Please read Daniel's 🧵
#note1hrf…3dhm

2025-10-05 01:40:50 CEST

- reply

Try `/model`. My default was 4.5 which I think they moved you to automatically if you had the default model enabled.
https://files.mastodon.social/media_attachments/files/115/318/535/726/210/346/original/658153ee1e7e54ae.png

2025-10-03 19:18:59 CEST

🆓 Wagtail Space is free and virtual next week.
🎟️ Sign up to support the event and drop by to learn something new.
https://wagtail.org/wagtail-space-2025/
#Python #Django #WagtailCMS

2025-10-02 05:17:36 CEST

- reply

LLMs are REALLY good at writing Django code 😉

2025-10-01 16:56:17 CEST

RE: https://mastodon.social/@djangonews/115299458225842327
Just a reminder that @npub1qsd…3jep is a great way to get the latest @npub1l0c…kt92 security releases and other details, often hours before they appear elsewhere.
#python #django #security
#note1tmr…wusn

2025-09-29 14:16:13 CEST

- reply

This is great. I was a big overwhelmed setting my default settings with Granian the last time I tried to configure it. I ended up punting knowing someone else would figure it out for me. I'd love to see settings you landed on that aren't covered in your article if you ever feel like sharing.

2025-09-26 16:18:26 CEST

It’s wild how **Django** has become an implicit standard in LLM benchmarks. Even when companies are not building in Django, they use it to measure AI code performance.
In **SWE-Bench Verified**, over **40% of the tasks come from the Django codebase** — but the benchmark focuses on tiny edits (median ~4 LOC), not the real complexity of maintaining production software.
👉 [Read the analysis](https://blog.nilenso.com/blog/2025/09/25/swe-benchmarks/)

2025-09-24 19:13:32 CEST

I'm not sure why, but my brain seems to work well with bending LLMs.
@npub1mjx…httg told me that he couldn't get the word "VOTE" to appear in an image, and it only took two suggestions to get it working. My other suggestion was to have it spell it out via "V" "O" "T" "E", which I have had to use before to get some signs printed for our office once, and it was confused by "revsys".
Anyways, enjoy:
https://wsvincent.com/chatgpt-dark-arts/

2025-09-21 17:54:04 CEST

- reply

This wasn't meant to be vague. Check the orange website or DHH’s recent X(twitter) posts or blog posts plus the recent Ruby Central debacle plus the last decade or going off the rails moments which always coincide with product launches. So ya. Over it.

2025-09-21 04:40:37 CEST

What a shame, Omarchy looked appealing, and I lost all desire to want to install it after the last few weeks. What a liability.

2025-09-20 21:21:47 CEST

- reply

Exactly this.
If we go back a year, Django was bigger than FastAPI. Skip forward to today, and it's 10x to 12x bigger. Python even grew 26% over last year (last stat was from June), but Django has plateaued in downloads.
We can overcome this, but we have to take it seriously, and start telling people what Django can do (and update our docs.)