Rijn (@RijnHartman) header
Rijn (@RijnHartman) profile picture

ex college basketball player and high jumper. into ancient civilizations, stoicism, science, power and how people think. building with AI most days

South Africa rn
Joined June 2012
Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)Photo by Rijn (@RijnHartman)
Rijn retweeted
Well isn’t that interesting 👀
BlackRock sold $541,000,000 in $BTC yesterday. The largest single day outflow since inception.
85
313
25
5,352
620,565
As a designer, what I love most about code and AI is regenerative design: designing and composing through code, following the structure I want. Why is it the most useful design tool for me? Once you’ve designed a visual universe, you can extend its life in many ways. It’s not a flat design you have to keep pushing forward yourself and redrawing every corner; the code runs the loop and keeps it alive. With a few small export scripts for formats like WebP, PNG, or SVG, you can carry that work across applications and keep it fresh and dynamic. Take the hero animation on lorediv.com. It’s built entirely in JavaScript and runs in a loop, so I can adjust speed, angles, perspective, size, transparency, text, labels, numbers of elements, colors, etc.. just by changing the code. That means I can create an animated version for a landing page or anywhere else, while also building a model behind the scenes that exports individual assets just as quickly. A design team can build a flexible library of visuals and reuse it across banners, social posts, landing pages, and more, creating consistent materials faster while adapting to each platform and keeping the brand fresh.
1
1
3
60
the same way i did before, there's just a lot more of it now clients pay me to build the thing and then a monthly to keep using it. about 3x the work is landing since august, so $800 isn't a hard bill to cover
Replying to @RijnHartman
How do you make money for roi?
1
77
i've got another sub. you're welcome for the ipo, anthropic
i'm adding a third $200 claude plan. that takes me to $800 a month on ai i had claude check my own usage first to see if i can justify it i'm using about 4x the tokens i did in august. about 3x the commits are landing, across twice as many projects. the last 30 days came to roughly $20k of api value on $600 of plans but this is what decided it. lately my claude accounts have been behind pace, meaning i'm burning through them faster than the week allows. the work isn't done when the limits get low, so i slow down i've been holding back and pacing myself so the third plan is so i can just confidently prompt and orchestrate agents to close things out, while i start new projects and new ideas, or audit what's left on client work, personal stuff, family business stuff, everything i'm juggling i feel like my head will be in a lot clearer space
2
5
177
marcus aurelius ran rome and still had to write himself a note to stop watching the neighbours we have the same problem, the neighbourhood is just the whole internet now
1
2
122
the rest of the line: "but only to what he does himself, that it may be just and pure" meditations 4.18, george long's translation
1
19
as a solo founder there's always someone on the timeline shipping more than me. none of them are building my stuff
18
since i moved to t3 code i'm orchestrating a lot with codex. 5.9B tokens this week on openai models it went from 11% of my tokens to 41% so yes, put it in the tab bar
We’re also working on a new mobile experience for ChatGPT! It’s easier to get to your favorite features and discover what you can do in ChatGPT. We’re starting with Pro users who have a dot or are eligible to create one. Your dot gets its own tab, and you can pin Codex, Projects, Images, and more to the tab bar. Some of you will see the new experience by default. You can turn it on or off anytime in Settings > New ChatGPT Design. Super excited to hear your feedback before we roll it out more widely!
1
2
260
i'm adding a third $200 claude plan. that takes me to $800 a month on ai i had claude check my own usage first to see if i can justify it i'm using about 4x the tokens i did in august. about 3x the commits are landing, across twice as many projects. the last 30 days came to roughly $20k of api value on $600 of plans but this is what decided it. lately my claude accounts have been behind pace, meaning i'm burning through them faster than the week allows. the work isn't done when the limits get low, so i slow down i've been holding back and pacing myself so the third plan is so i can just confidently prompt and orchestrate agents to close things out, while i start new projects and new ideas, or audit what's left on client work, personal stuff, family business stuff, everything i'm juggling i feel like my head will be in a lot clearer space
5
1
3
338
my upload speed in magaliesburg today was 0.3 mb a second. it took almost 3 hours to publish a 158 mb mac app whatever the answer on starlink is, i'd like it soon
1
1
194
anthropic just raised $81 billion can i use more than 50% of my weekly limit on fable now
7
351
so we're all making our own gta now
Introducing Genie 4 in the Gemini API + @GoogleAIStudio, our first world model for developers to build with and create new world model experiences 🔮 Try it out: ai.studio/genie
4
291
Rijn retweeted
fable 5.2 is good at swiftui animation asked it to upgrade this drag interaction in latch last night. the first version took it 16 minutes. this is the second, from this morning the before is in the post below
asked fable 5.2 to upgrade this drag animation in latch and make it cool instead of orchestrating, it said "this is mine to do, and i'd like to" here's the before. when you drag something towards the notch on your mac, a morphing thing comes out, follows your cursor, and then latch opens up will post the after
4
1
8
388
Rijn retweeted
my family's 6 pharmacies run on whatsapp groups and 13 google calendars i'm putting all of it into one app, and payroll too chat every branch gets its own channel announcements have a "got it" button so you see who still hasn't read it voice notes, photos and pdfs like whatsapp roster staff pick their off days in the app and it tells them when a day is already short it builds the month for every branch around the rules. a pharmacist on whenever the doors are open, max 45 hours a week, as many hands as possible on month-end and sassa pension days someone's sick, they tap one button, and the app shows who can cover. or you book a locum payroll tax and uif worked out for everyone nobody gets paid until the roster, the clock-ins and the till agree leave builds up by itself payslips go straight to each person's phone the monthly tax return and the bank payment file are ready to upload it's in testing with made-up staff. the family goes on next then my clients. it's modules, so each business only switches on what it needs. we're also building tills and pharmacy systems for other businesses, and those can plug in later so eventually an owner runs the whole business from one app
1
4
232
building a till for a pub where the owner runs the bar alone found a bug today that only happens because he's alone. customer has a R100 tab, pays R95, he gives 5% off, bill hits R0. the till saw R0, decided it was a freebie and wanted a second manager to sign it off he's the only one there. so he couldn't close the bill and couldn't cash up fixed now. when you're working alone you approve it yourself and your name, the time and the reason go on record
1
74
fable 5.2 is good at swiftui animation asked it to upgrade this drag interaction in latch last night. the first version took it 16 minutes. this is the second, from this morning the before is in the post below
asked fable 5.2 to upgrade this drag animation in latch and make it cool instead of orchestrating, it said "this is mine to do, and i'd like to" here's the before. when you drag something towards the notch on your mac, a morphing thing comes out, follows your cursor, and then latch opens up will post the after
4
1
8
388
my family's 6 pharmacies run on whatsapp groups and 13 google calendars i'm putting all of it into one app, and payroll too chat every branch gets its own channel announcements have a "got it" button so you see who still hasn't read it voice notes, photos and pdfs like whatsapp roster staff pick their off days in the app and it tells them when a day is already short it builds the month for every branch around the rules. a pharmacist on whenever the doors are open, max 45 hours a week, as many hands as possible on month-end and sassa pension days someone's sick, they tap one button, and the app shows who can cover. or you book a locum payroll tax and uif worked out for everyone nobody gets paid until the roster, the clock-ins and the till agree leave builds up by itself payslips go straight to each person's phone the monthly tax return and the bank payment file are ready to upload it's in testing with made-up staff. the family goes on next then my clients. it's modules, so each business only switches on what it needs. we're also building tills and pharmacy systems for other businesses, and those can plug in later so eventually an owner runs the whole business from one app
1
4
232
asked fable 5.2 to upgrade this drag animation in latch and make it cool instead of orchestrating, it said "this is mine to do, and i'd like to" here's the before. when you drag something towards the notch on your mac, a morphing thing comes out, follows your cursor, and then latch opens up will post the after
1
1
477
gemini 4 neon just came out so i put it next to claude and gpt-6.2 it's good. 61 on the artificial analysis index, only fable 5.2 and astra are above it, and it's $3 in $15 out but look at what a task actually costs, not the list price sol 6.2 and sonnet 5.5 cost exactly the same per token. on the same test sol came to about $1 a task and sonnet about $10 neon came to $6.44 a task. opus 5.5, which scores lower, was $10.04 fable is the smartest of all of them and it's $26.54 a task, 41 minutes each so the cheap looking model isn't always the cheap one. some of them just burn way more tokens to get to the answer
Google continues to accelerate its release cadence with Gemini 4 Neon, delivering a small Intelligence gain over Gemini 4 Carbon at a ~17% lower Cost per Task @GoogleDeepMind has released Gemini 4 Neon, its fourth model in just seven weeks. Neon (xhigh) scores 61 on the Artificial Analysis Intelligence Index, 1 point ahead of Gemini 4 Carbon (high) and 3rd overall, behind GPT-6.2 Astra (max) and Claude Fable 5.2 (max). At high effort, Neon matches Carbon (high) at less than half the Cost per Task using Carbon’s standard (non-discounted) pricing. Key takeaways: ➤ Slight improvement in Intelligence for lower Cost and Time per Task: Neon (xhigh) scores 61 on the Intelligence Index, 1 point ahead of Carbon (high), with small improvements in agentic knowledge work, visual reasoning and coding. It costs $6.44 per task, ~17% less than Carbon ($7.76 at non-discounted pricing), and takes ~21 minutes per task, ~40% less than Carbon (~36 minutes). Neon uses ~15% more output tokens per task, so the lower cost comes purely from its 25% lower token prices ($3/$15 vs $4/$20 per 1M input/output tokens), and the lower time from its higher output speed (~110 vs ~57 output tokens per second) ➤ Competitive with Claude Opus 5.5 and Claude Fable 5.2 on coding agents: With Antigravity CLI, Neon (xhigh) scores 59 on the Artificial Analysis Coding Agent Index, 3 points ahead of Carbon (high, 55). It scores in line with Claude Code with Claude Fable 5.2 (xhigh, 58) and Claude Opus 5.5 (max, 58) at a much lower Cost per Task. Gemini 4 Neon in Antigravity has an average Cost per Task of $6.86, compared to $12.61 for Claude Opus 5.5 (max) and $15.47 for Claude Fable 5.2 (xhigh). Neon is available in Antigravity CLI for Pro and Ultra subscribers ➤ Significant improvements in visual reasoning and output presentation quality: In GDP.pdf, Gemini 4 Neon (high) scores 33.4%, 5 points above Carbon (high, 28.4%) and in line with the highest scores measured, while Neon (xhigh) scores 28.0%. On AA-Briefcase, Neon (xhigh) scores 1700 Elo overall and leads on Presentation Quality with 1771 Elo, 45 points ahead of the previous leader Claude Sonnet 5.5 (max, 1726). On Google’s Multimodal Knowledge Agent evaluation (MKA), Neon reaches a Climb WSR of 56%, up 2 points from Carbon and on par with the leader, Claude Fable 5.2 (max, 57%) Key model details: ➤ Context Window: 1M tokens ➤ Multimodality: Text, image, video, and speech input, with text output ➤ Availability: Gemini 4 Neon is publicly available at launch through Google’s API. Pro and Ultra subscribers can also access it in Antigravity and the Gemini app ➤ Effort levels: Gemini 4 Neon offers four effort levels (low, medium, high, and xhigh) - xhigh was introduced to Gemini models in the recent Hydrogen release ➤ Pricing: $3/$15 per 1M input/output tokens, 50% above Gemini 4 Carbon’s currently discounted price of $2/$10 (which ends on November 30), and 25% below its standard $4/$20. Cached input tokens receive the same 95% discount ($0.15)
4
170
x's ux on desktop can be so much better give one person claude and a weekend
1
4
120