Proud founder moment 🙌🙏
Today we're launching the first product from Tunelo:
GarageManager — a workshop management tool built for South African auto repair shops.
We didn't set out to build SaaS. Tunelo started as a consulting operation, helping SA businesses figure out where the real friction in their workflows lived.
5
9
393
Your future self will thank you.
When you feel like giving up, then it’s time to push even harder. That breakthrough will come soon enough!!
Received the first proper user review for GarageManager after many attempts requesting feedback, this feels better than a paying sub 🙌 2 in a day!😎
Trimm 🥷 retweeted
My new favourite words “US only” 😬😔🙂↔️😖
We need to push for South African alternatives.
Laptop closed. Buzzy’s still buzzing.
Meet remote agents in Buzz. Pick a model, connect your tools, and put them to work in the cloud. Check in from your phone.
Lucas shows you how to get started ↓
US only for now. buzz.xyz
Phew, ChatGPT for Accounting is now available.
Our new Accounting plugin and partner plugins help you close your books with less manual work for tasks like:
- Month-end reports
- Bank reconciliation
- Data entry
- Audit prep
Put ChatGPT to work with tools like @QuickBooks, @Xero, Oracle @NetSuite , @BILLcom, @FreshBooks, @RilletHQ, and @Meet_campfire.
Try it here: openai.com/business/solution…
Trimm 🥷 retweeted
Can sit and moan or make the best of it. I’ll be watching the match at the hotel. Not keen to watch in on my own. So lets give it a go and hope I get the “tech” sorted = TV on, My Laptop camera on, and you’re welcome to join me. Sign up here (adults 18+): watchwithrassie.com. This is personal, not official.
So this happened 🎉
Next we want to look at NVIDIA's speech and document tools, for South African accents and Afrikaans, and for scanned gazettes and tender packs that ordinary text extraction can't read.
Thanks to the NVIDIA Inception team for having us. Back to building 👨💻🏗️
#NVIDIAInception #AI #SouthAfrica #Startups
Tunelo has been accepted into NVIDIA Inception, NVIDIA's program for startups building with AI.
We build RepoLens, GarageManager and Industrial Hub for South African businesses. Back to building.
tunelo.co.za
#NVIDIAInception #AI #SouthAfrica
Trimm 🥷 retweeted
I didn't say it, Grok did. What are you waiting for? Sign up to versionone.co.za
Replying to @DistinctLee
Version One is a focused, purpose-built showcase for South African builders shipping early versions and finding first local users. With 150+ releases and steady weekly activity since launch, it fills a clear niche that broader platforms like Product Hunt or Indie Hackers do not prioritise. Whether it is the single best depends on the goal, but for Mzansi early visibility it is currently one of the strongest dedicated homes available.
Trimm 🥷 retweeted
Tomorrow morning I'm on @Powerfm987 Breakfast talking about how AI is helping developers build and ship faster, and about ElectAI with @LeboPhokela.
Tune in at 06:40 on 98.7 📻 👋🏾👋🏾
Trimm 🥷 retweeted
1/11 Rules of engagement for the 21-day builder challenge:
Post about your product at least twice a day. That's 42 posts over 21 days. You can post more. Use whichever platforms you like. Text, screenshots, demos or video. The aim is to keep showing your work.
Trimm 🥷 retweeted
I'm looking for more people on X who build things around a full time job.
Not necessarily startups. Could be a research project, a weird hardware project, a little SaaS, or something you're teaching yourself.
I'm particularly interested in people who only get a few hours a week to work on their ideas.
What are you building?
Trimm 🥷 retweeted
Tunelo's first app is live on the App Store.
RepoLens turns South Africa's sheriff auction notices from the Government Gazette into clear listings: auction date, venue, sheriff and trip cost. All nine provinces.
Free on iPhone. Android soon.
Here's 4 steps to run an uncensored AI on your own GPU for $0:
I run a local abliterated Gemma 4 for anything I don't want leaving my machine
By the end, you'll know how to:
- run a model which has no refuses
- plug it into Claude Code
- do it all for $0, no API key, no account
So, let's go step by step:
1. Install Ollama
curl -fsSL ollama.com/install.sh | sh
you want a recent version, older builds don't speak the Anthropic API and don't support Gemma 4
2. Pull the model
ollama pull igorls/gemma-4-12B-it-heretic-GGUF
the biggest mistake people make here is grabbing a random "uncensored" tag with no quality numbers
this one is abliterated with Heretic. refusals went 99/100 → 0/100 at a KL divergence of 0.028 from the original, so you keep almost all of the base model's capability
one thing: it's still Google's Gemma license underneath, fine for personal use, but read the terms before you build a product on it
3. Turn off thinking (the part every other guide skips)
Gemma 4 is a hybrid thinking model and the abliteration targets non-thinking mode
leave thinking on and the reasoning block eats your token budget and the answer gets cut off
ollama run igorls/gemma-4-12B-it-heretic-GGUF
/set nothink
if you're calling the API instead, add "think": false to the request
4. Plug it into Claude Code (OPTIONAL)
ollama launch claude --model igorls/gemma-4-12B-it-heretic-GGUF
or manually:
export ANTHROPIC_BASE_URL=http://localhost:11434
export ANTHROPIC_AUTH_TOKEN=ollama
claude --model igorls/gemma-4-12B-it-heretic-GGUF
two gotchas:
- bump the context first. Ollama defaults to 4k and Claude Code's system prompt alone is bigger than that, so the model never sees its tool instructions. restart the server with OLLAMA_CONTEXT_LENGTH=32768 ollama serve (setting it in your shell does nothing if Ollama is already running as a service). this needs more VRAM than the 8GB below
- Gemma 4 tool calling is still rough, you should expect some failed edits, imo for serious agentic work use qwen3.5 or glm-4.7-flash instead
What you need:
- 8-12GB VRAM for Q4 (12B)
- ~8GB of disk for Q4, ~13GB for Q8
- more of both if you want a big context
honest take:
- "abliterated" means refusals removed, not quality added. it's still a 12B
- fewer refusals ≠ smarter. for real coding work I still use cloud models
- you own the weights, your prompts never leave the box. that's the actual win
*IMPORTANT: don't use it for anything illegal or harmful. that's on you, not the model
now i'm collecting a list of every uncensored model i'm adding to my local stack and which tasks each one replaces for me
also planning a big "Uncensored 101" article, still figuring out all the angles
save it so you don't lose it
First menu is out!
Lets have a look at whats cooking! 🔥🔥🔥
Cooking or Fried?! Episode 1 👨🍳
We crossed 100 builders on Version One 🎉 To celebrate, Gordon "Lee" Ramsey is in the kitchen 🤣
On the menu: @HondeAI by @tawanike and @sitespeakai by @hermanschutte.
Watch the full episode. What should I taste next?
1
2
47
Everyone in Mzansi should be showcasing their projects here. Builders pull up! 👩💻👨💻
versionone.co.za/join?invite…
Trimm 🥷 retweeted
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org/
- Excellent recs without any logins (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.




















