a page titled "Know your crowd" with 3 other pages showing different statistic questions like "How long have you been programming" or "light mode vs dark" mode. The statistics are filled with colorful, funny stickers
The first edition of our @rebased.wtf meetup was amazing - we were SO glad about all the positive feedback 💖 We're happy to announce the next one on 20. May. Let's have fun with tech. See more on rebased.wtf
Together with a few friends (
@bitsbysigrid.bsky.social @mxmlnznr.bsky.social ) I'm starting a new tech meetup in Vienna called @rebased.wtf
Check it out and register on Luma!
Neues Projekt, check it out: mietmonitor.at
Danke @momentum-institut.at und @lisaduschek.bsky.social für die Zusammenarbeit 🤝
If I ever win the lottery, I won't tell anyone.
But there will be signs...
This year might have broken my will to try any startup DB offerings tbh
I just deleted my remote Cloudflare D1 db because the command doesn't have a remote/local flag (like other commands).
And now the sql dump I got from Cloudflare can't be imported because of a syntax error 🙃
2023 was the craziest year of my life:
I started the year with a new job and was fired after 3 weeks, got involved in politics, made apps for sex positive parties, was invited to consult a minister on a new law and much more.
Read the post: www.chrcit.com/articles/202...
A pie chart titled "Model Usage Distribution" illustrating the distribution of usage among different AI models. The chart shows a large purple segment labeled "claude-3.5-sonnet" with a value of 204, indicating the predominant model used. A small green segment represents "claude-3-opus" with a value of 5, and a tiny yellow segment indicates "o1-mini" with a value of 1. The overall design emphasizes the significant difference in usage between the models.
This is the model distribution lol
You can get the history by clicking on pricing details here: cursor.com/settings
And I asked Claude to visualize it with the experimental feature flag
I might have to write a scraper which gets me this data from the page every day.
A bar chart titled "Actions Timeline" displaying the last 500 cursor actions over several days. The x-axis represents dates and times from November 17 to November 20, with specific time markers. The y-axis shows the number of actions, reaching up to 80. Different colors represent distinct actions: purple for "fastApply," green for "composer," yellow for "warmComposer," orange for "cmdK," light blue for "chat," and dark blue for "contextChat." The chart highlights peaks in activity, particularly on November 17 at 16:00 and November 19 at 18:00.
My last 500 cursor actions
Just multi channel posting to here and Twitter would be great.
If you add analytics on top I'm super happy!
A screenshot of a code analysis interface showing a timeline of actions with date/time formatting on the x-axis. Below, there is a debug card displaying raw timestamps and their parsed values using different methods. The output indicates that the manual parsing method works correctly, while one timestamp is marked as NaN (not a number), indicating a parsing error.
If Claude is ever stuck on a small issue tell it to try multiple variations.
Here it was stuck with some basic date parsing so I just told it to output me variations.
Instead it REPLed it and fixed the issue.
Hades 100%, great gameplay, art direction and story telling!
I played it so much that I saw the animations when I closed my eyes when going to sleep 😂
8. Bluesky is an open ecosystem with an API that is guaranteed to always be open to developers.
Developers, we're eager to see what you create! Our API docs are here. docs.bsky.app
The openness of the platform is really exciting – I already have it on my list for next year to build something with the API!
The amount of cool projects that pop up around Bsky is amazing. That's when your social network is hackable I guess
bsky-follow-finder.theo.io
I made a little starter kit for future side projects.
Stack: TS, Remix+React, Tailwind, pnpm, Plausible
Will probably fork it to create starters for Payload, Sanity and Drizzle.
github.com/madebyarthou...
Twitter wasn’t super unprofitable and had decent revenue at 5B/y.
Running the company 30% leaner + revenue share + pro memberships and it could have worked out.
Elon just tanked advertiser trust for no reason.
Honestly, running it as a leaner Twitter should be possible.
Ad revenue and a decent membership system would bring in enough to sustain a sizeable team.
Only Elon could have fumbled this so hard.
👋