been a developer for a long time (learned to code when I was 12). I like to build dev tools for myself, explore ideas and experiment with crazy ideas.
Joined November 2021
- Tweets7.1K
- Following305
- Followers2.4K
- Likes12.9K
So I was talking to someone I met while trick or treating with my kids that I should do game programming in Rust, Python or Ruby.
I honestly don’t know why I would want to do it in Python or Ruby but I guess? 🤷🏻♀️
Going out with my kids soon. Get to see inside me to know what kind of person I am.
(Had to repost with version because the other one had something that I shouldn’t show.)
Today someone wrote "20 ttl" as a shorthand for "total" at my workout. I read it as 20 ++1 so I started at 20 and kept adding one rep every time. It got rough after a while.
This is what programming does to your brain.
So I woke up at 4am. Showered. Got in my workout clothes. Walk out of my bedroom and what do I find? My two kids, in their costumes. I tell them "No way, you need to go back to sleep."
Then they joyfully tell me "We've been awake since 2am!" I just walked away and went to work out after that.
It's going to be a long day after they get out of school.
I love working out so much I’m going to get up at 4am tomorrow to do it.
This is what life has come to.
I feel like such a mom when I’m in workout clothes running my kids around. I don’t wear leggings that cling to my butt but at this rate, I’d do it to complete the vibes.
My husband is hurt pretty bad so had to step up my mom game.
My kid let me ride her rollercoaster in Minecraft last night. She's so creative.
Also that tiny arrow right above the control bar is my waypoint. Totally works. I love java.
You can’t tell me you don’t do this with your friendly neighborhood LLM as well. Thanks jippity. Good rocking out.
Inspired by something Prime posted. Don’t judge me because I’m creative.
I’ve been baking more which is funny because I don’t eat the desserts I make. But it’s fun to make treats to delight my family with.
This is my first pumpkin dump cake. Is it good? No idea. :)
It’s funny how there are people panicking over Tailscale and meanwhile letting AI read their api keys and use their computer.
Not saying it’s wrong. But it’s funny.
Said hello to Java again and wrote a Minecraft mod over the last few days. It adds a minimap and a full screen map that show everyone on the server.
My kid is the red dot. Next up is waypoints.
(Also my kid had such a blast testing this with me. Such joy.)
codegirl007 retweeted
Uncensored DeepSeek V4.1 Flash for almost nothing
> 1M context
> Zero data retention
> $0.12 / 1M input
Just sign in with GitHub or Google
larprouter.com
codegirl007 retweeted
BEST Models for 96gb of VRAM / RAM
I'm specifically focused on models with good browser use capabilities/good coding/decent vision.
I recommend either:
- sitegeist: allows the model to do anything codex browser use can do in the background github.com/0xsero/sitegeist
- cua github.com/trycua/cua (computer/browser use)
--------------
Qwen-3.8-next-flash:
- score 40 on AA
- Gemini-3.8-flash levels / GPT-6-Luna max level
- very fast
- good at browser/computer use
- need an extra 50gb of nvme/ssd
It's a good model, hosting it is a slight bit more weird but again any free model on Opencode can get it running quick.
I recommend:
huggingface.co/nvidia/Qwen3.…
---------------
Qwen-3.8-27b:
- Score: 34 on AA
- Opus-4.6 perf
- speculative decoding on
- medium speed
- overthinks
Very smart, capable of doing pretty much all computer use tasks, outputs good UIs/assets, and generally performs well.
Here's a very optimised version, otherwise any 8bit from Huggingface will do.
github.com/kacper-daftcode/q…
Opencode free/claude/codex can setup the right one with all the details.
huggingface.co/turboderp/Qwe…
-------------
Ornith-1.5-122B:
- 81.2% on terminal-bench-2.1 matching ds4.1
- also pretty fast
Post train of Qwen, pushes it's performance a fair bit
huggingface.co/Mia-AiLab/Orn… <-- if you give this to codex/claude/kimi whatever they'll get it running fast
------------------
Nanbeidge:
- Dense
- Very smart
- Very slow
Worth looking into this more, but i'd recommend the qwens
huggingface.co/Nanbeige/Nanb…
============
If you can get one of these models running, please consider asking it to push it's recipe up to github.com/sybil-solutions/l…
Nah. My kids are allowed one piece of candy a day from Halloween if they remember to toss their wrappers in the trash. After a couple days to a week, they forget about it.
I don’t think there’s been a year where I’ve let them go all out on eating candy.
A biological attack is planned for Saturday
600 million pounds of sugar laced with synthetic chemicals will enter the bodies of our children.
Wreaking metabolic havoc, injuring vascular architecture, acidifying oral lining, releasing bacterial toxins, and blunting neural reward pathways.
Instructing that gluttonous consumption of poison is celebratory.
Ugh YouTube, why did you change the layout of the mobile app? Now I’m forever going to hit that + button when I just want to see my subs.
Going to start posting weird videos on YouTube because of this.
I’m going to play with this later today. Gotta try it.
Now live: abliterated-model-large-flash (our version of DeepSeek-V4.1-Flash-1017). It is a cheaper alternative to our GLM-5.3-based model for product owners and defenders who need access to the most cyber-capable models.
- 1M context
- Vision
- Zero input/output retention or training
We believe paying customers should be in control, and over the past few months, leading experts and major brands have adopted our models.
abliteration.ai