Recent content on Mitja Martini
When I was young, I thought of senior engineers, consultants, etc. as experts that are more effective. My assumption was that seniors work harder to get there and be senior. Now that I’m older, I realize: Being efficient is as important and probably more distintctive than being effective: I need to be calm, save time, and work efficiently to achieve the goals in a relaxed way, as I…
Today I learned about Rashid’s solution to export ChatGPT conversations to Markdown or even PDF if you like. I prefer Markdown as I can use it to continue the conversation with Claude for example, or change the content and use it in other contexts. You don’t need to install anything because it’s just a JavaScript snippet. You open the conversation in a browser, select anything in…
The recent RAM price hikes have pushed GPU prices up as well. The only systems that have not yet been affected to the same extent are Macs and high-end GPUs (RTX 6000 Pro and above). However, I would already classify GPUs as out of reach: running coding models with sufficiently large context windows would require one or two RTX 6000 Pro cards, or three to six RTX 5090s. Let’s take a look at Macs…
Christian Posta from Solo.io has written an interesting Deep Dive into MCP and A2A Attack Vectors for AI Agents . Here is a shortlist. There are certainly more attack vectors, and the mitigations are a start, but certainly not perfect. Attack Vector How it works Mitigation Naming Attacks Attackers create look-alike names or typosquatted services that trick the AI into picking a malicious resource…
In case you haven’t tried vibe coding, recently, you should probably try it (again). Their performance has increased quite a bit thanks to recent models like Opus 4.5. Here are some recent takes on vibe coding: Everyone’s talking about vibe coding without looking at code. I was skeptical. I decided to give it a shot on a challenging problem and was blown away by what I could accomplish…
Here’s a trick from the 80s that’s still useful, today: Van Halen’s live shows were potentially dangerous. To make sure the local crew read the rooster, they asked for a candy bowl with M&Ms and added: Warning: Absolutely no brown ones. The Van Halen Test via Left hand writes One day in 1982, they trashed their backstage area because David Lee Roth found a brown M&M. With coding…
So far, my web apps run on a Dokku server. I haven’t tried Vercel or Fly because I didn’t want to deal with complex pricing models that incur more costs with every additional project. Dokku works like Heroku: dokku apps:create myapp dokku postgres:create myapp-db dokku postgres:link myapp-db myapp dokku config:set myapp APP_SECRET = ' $( openssl rand -base64 48 ) ' dokku git:from-image…
Vibe scripting, for me, is when I develop small tools for myself with the help of coding agents. It works extremely well, especially for command-line tools. I recently told a tax advisor about it. He was interested and wanted to see how it works. So I developed a small example on the spot: a VAT calculator. Not the best example, but I couldn’t think of anything better on short notice. It…
Terms and conditions
OpenAI just releasesd MCP connectors and ChatGPT developer mode beta. In this post, I describe the process of connecting MCP servers to ChatGPT, show how they look and feel right now in a chat session and give an overview of their current limitations. Activating developer mode # MCP connectors can only be created and edited in developer mode which can look a bit scary: Note: This article is…
Development Containers or just “devcontainers” add a layer of security, simplify developer onboarding enables developing in parallel with isolated environments. For me, Devcontainers are a great addition to an AI Engineer’s toolbox, even though I don’t use them day-to-day. I have setup Devcontainers with Claude Code using VS Code and the devcontainers cli, adapted it to a…
About me
Privacy statement
I’ll help you turn prompts into reliable AI features and apps.
In this article, I document some non-technical things I’ve learned migrating my blog from Jekyll to Hugo . The gist is to keep it as simple as possible :
Notes of a whirlwind intro to evals for voice agents by Kwindla and swyx
After reading a nice post via HN about On Stepping Back , I habitually scrolled through the comments and found this nugget by evrimoztamur : Interacting with LLM coding tools is much like playing a slot machine, it grabs and chokeholds your gambling instincts. You’re rolling dice for the perfect result without much thought. This probably applies to many systems that use Generative AI,…
Pipecat Cloud is located in the US. Is its latency ok for voice agents for EU Users.
A somewhat hidden feature of Ollama is K/V Cache quantization. This is relevant for local AI as it reduces memory consumption, especially for small LLMs with large context windows. This post describes how to activate K/V Cache in Ollama and gives an overview of its benefits, drawbacks and use cases. Activating K/V cache quantization in Ollama # K/V Cache quantization in Ollama is not on by…
Here are my notes on the session about deploying voice agents to production which is part of the Voice Agents Course . The course if held by kwindla und swyx, the CTO und an investor of Daily.co, a WebRTC und Voice AI infrastructure provider. Some of their recommendations might be predisposed. I still state them as is as I trust them and because I don’t have enough experience with voice…
I’m so happy to be part of the Voice Agents Course by Kwindla and swyx. Yesterday, Kwindla kicked it off with an overview of the voice AI landscape. The pace, insights, and questions from the audience were just great. Here are my personal notes, probably incomplete and maybe not always correct. For a more authoritative overview of the Voice AI landscape, check out their free online book…
In The Batch 300 , Andrew Ng shared some insights about the importance of speed for startups and how to move fast as a startup. I think they apply well to ideas for a fast solopreneur: Focus on one idea, Code prototypes intuitively, Be creative about getting user feedback quickly (it doesn’t have to scale), Pivot quickly but stay within your domain, Use a KISS stack you know, and Learn AI…
How to Install Docker and Docker Compose on Raspberry Pi OS (rootful and rootless).
AI Apps are often expected to be realtime. On the web, realtime communication can be implemented with WebSockets. I’ve started with WebSockets to create chatbots and other live-updated interfaces, but then switched to SSE and now mostly follow these rules of thumbs: Use WebSockets for server-to-server communication. Use SSE for server-to-client communication. If you still want or need to use…
A look at the NVIDIA RTX 5090 specs for local LLM inference.
The ChatGPT search feature can also search and use information from intranet websites.
An interesting statement by Yann LeCun.
This article describes how to deploy a static site on Dokku, including activating Let’s encrypt signed certificates for the domain. Configure DNS # Register the domain and point it to the Dokku server. I usually set these records in the domain’s DNS zone: A @ 7200 $DOKKU_SERVER_IPv4 A * 7200 $DOKKU_SERVER_IPv4 CNAME www 7200 $APP_DOMAIN AAAA @ 7200 $DOKKU_SERVER_IPv6 AAAA * 7200…
A quick tip to make working a bit easier.
A quick tip to make your life a bit easier with VSCode and Python.
In this post I’ll share the steps I did to deploy a SaaS Pegasus bootstrapped Django app (Scriv from the SaaS Pegasus marketplace to be precise) with Celery, Redis and Postgres on a Dokku host. There were some hickups along the way, but I believe, when you go step-by-step, it is quite straightforward.
A quick note about the Furo theme for the Sphinx documentation system.
A workaround for transferring models to air-gapped Ollama instances.
A tutorial and video about installing and using Ollama and OpenWebUI on Windows.
How to summarize texts with LLMs that are too large for their context window.
Example Excel workbooks for a book about using OpenAI models in Excel Formulas.
Practical use cases of ChatGPT’s / GPT-4’s image input feature.
What are small LLMs good at and how to get good results from small LLMs.
How to install and use httpx.
Microsoft keeps adding new exciting features to Excel all the time. Here is an overview of new features added in the first quarter of 2022. For me, TEXTSPLIT is the best new Excel feature of this quarter. With it you can split text that is eg. formatted as CSV into multiple rows and columns. This is extremely helpful if you want to use ChatGPT in Excel. I base this on the Excel Blog. Please drop…