
My heat pump is a Kronoterm: Hydro S indoor unit, Adapt 0416 outdoor unit, TERMOTRONIC 3000 controller. Good hardware. The software situation is that the only official way to see or change anything remotely is the Kronoterm cloud app, which looks and behaves like a side project from 2012, because it is one. Slow, missing half the interesting registers, and every command round-trips through Elektrina's servers to reach a box two meters from my router.

I wanted full local control in Home Assistant. The existing community integrations all wrap the same cloud API, so they inherit its limits. The other option was to remove the cloud from the equation entirely. That's what I did, with Kimi K3 running inside the pi coding agent doing most of the work: we reverse-engineered the module's protocol, impersonated the cloud endpoint, and finished with a proper local integration. Here's how it went.

## Recon

The TTEH-WEB module sits on the LAN with exactly one open port: 80. The web UI ships with admin:admin (yes, really) and offers no local automation API. So the interesting question is what it talks to. Watching DNS: a persistent WebSocket over TLS to wd.elektrina.si.

Then it gets good. The handshake negotiates SSLv3 with RC4-MD5. In 2026. The certificate is 1024-bit RSA issued by a private Elektrina CA. And the module does not pin that CA. No pinning means the module will happily do its ancient handshake with anyone presenting a plausible certificate, which means I can be the cloud.

## The interception box

Two constraints: nothing persistent may change on the heat pump itself, and the intercept has to be always-on. I have a UniFi network and the UniFi Cloud Gateway is Debian with root over SSH and already sits in the path, so the first version went there: an alias IP on the LAN bridge, one iptables REDIRECT rule, one dnsmasq line pointing the cloud hostname at it.

That later moved to a tiny Proxmox LXC, and the dnsmasq hack was replaced by a proper UniFi static DNS record created through the Network API. I gave my UniFi credentials to the agent and it configured it unassisted. The gateway is now back to completely stock config.

## Talking 1998

Python's ssl module hasn't spoken SSLv3 in years, so the agent vendored tlslite-ng (still does) plus its dependencies; tlslite needs Python 3.11 or older because it builds on asyncore. Two more quirks: the module rejects 2048-bit certificates and insists on 1024, and the WebSocket runs a binary subprotocol called ttehweb_bin.

Phase one was a relay: terminate the module's connection locally, open a second one upstream to the real cloud, log everything. That captured full config syncs and steady-state traffic.

## The protocol

Frames come in a handful of types: 0x28 status updates, 0x29 bulk dumps, 0x0b and 0x1f config pushes. The fun bug is that endianness is mixed: 0x28 and 0x0b are big-endian, the 0x29 dumps are little-endian. This is exactly the kind of detail that eats an afternoon when you're decoding by hand. The agent ground through the captures, formed hypotheses, and tested them against live values (compressor on, 800 W electrical in, 5 kW thermal out, COP around 6, hot water at 49 C) until everything lined up.

The register map came from merging three sources: an official Kronoterm BMS/Modbus document, community register maps, and what we observed on the wire. 1140 registers total, about 180 documented, 67 writable.

Phase two replaced the relay with a full local emulator. The module connects, does its sync, and can't tell the difference. The official app now shows my heat pump as offline. That is not a bug; that is the goal. However, if I want the heat pump to start talking to Kronoterm again we can set up the upstream connection again.

## Home Assistant

On top of the emulator's small HTTP API, the agent wrote a real integration: config flow, polling coordinator, 42 entities. Climate entity for the floor loop, water heater for hot water, electrical power, thermal capacity, COP, SCOP, pressure, energy counters, switches for system/vacation/disinfection, heating curve points, the works. Published as a HACS-installable repo: [github.com/jernejstrasner/ha-kronoterm](https://github.com/jernejstrasner/ha-kronoterm).

![The Home Assistant dashboard on the local integration: status, efficiency, 24-hour graphs for temperatures, power, COP and energy](/images/2026-09-06/ha-dashboard.png)

The end result on my dashboard. Everything on it comes from the local integration; none of it touches Elektrina's servers.

Best detective moment: the documented energy counter registers read zero on my firmware. The real lifetime counters turned out to be two undocumented registers, confirmed because their ratio is 4.18, exactly the unit's reported SCOP. 10,751 kWh electrical in, 44,958 kWh thermal out over its life so far.

## Rollback

Nothing was ever written to the heat pump's own configuration. Full rollback is deleting one DNS record and rebooting the module; it re-registers with the real cloud within a minute. That property mattered to me more than any single feature. I definitely want to avoid any serviceablity issues down the line. It's my house's heating we're talking about.

## What the agent was actually good at

The grinding. Frame decoding, endianness hypotheses, register cross-referencing, writing a full HA integration with correct entity semantics, driving the UniFi and Home Assistant APIs directly instead of asking me to click through interfaces. Long captures went in, structured analysis came out, and it held the whole project state (register maps, frame formats, which services run where) without losing track. It also made classic mistakes: it pkilled its own shell once (the pattern matched its own command line), and Home Assistant's module caching made one "reload" look like my fix didn't work when it just wasn't loaded. Both got diagnosed and fixed without me touching the keyboard.

Cost and behavior, since people ask about both: the whole session ran about $20 in tokens. The same volume of work on Claude would cost roughly 4x that, at lower speed. And there was no refusal theater: this project is literally impersonating a vendor's cloud endpoint and intercepting its TLS traffic, and the model just did the work on my own hardware without a lecture. Anyone who has had an Anthropic model explain why it can't possibly help them modify their own equipment knows the difference.

The honest summary: a weekend-plus project compressed into an afternoon, and I spent that afternoon steering and reviewing instead of reading protocol documents. None of this is hard in the intellectual sense; it's tedious. Tedious is what agents are for.

Repo: [github.com/jernejstrasner/ha-kronoterm](https://github.com/jernejstrasner/ha-kronoterm) (the actual proxy service is closed-source for now)

If you've done the same with a Modbus-only unit, or you think MITM-ing your own appliances is a step too far, I want to hear it.

---

A note on the tooling, since it matters for how you read the above: I'm a founder at sference, where we host Kimi K3 ($3 / $15 / $0.45 per million tokens, [app.sference.com](https://app.sference.com)), and the agent doing the work was pi. I picked K3 for this because long-horizon grunt work is where it earns its keep: multi-hundred-line binary captures in, patient hypothesis testing, no drama. This post is the dogfood report.


My heat pump's cloud app was stuck in 2012, so I reverse-engineered the protocol and cut it out


I've been (vibe) coding a lot lately. I experimented with a few infrastructure projects (logging system, LLM gateway, tool calling permission management, etc.) but one that I found real passion around is an AI workout manager and coach. I built this one in parallel with the other, more "professional", projects out of frustration with other products that I've tried.

In the end I resorted to using Apple Notes and ChatGPT in order to plan out my training sessions. It was ok for a bit but quickly became too much to effectively manage and ChatGPT started losing context. Ok, I'm an engineer, we're in the vibe coding era, I can "fix" this.

![Calendar view](/images/2026-02-03/calendar.png)

# Background

I've been active my whole life, be it sports or even just workouts to stay healthy. As a high-schooler I trained rowing a lot, to the point where it got me to the brink of "do I go professional or not?". I had good success locally as well as at the European U23 level. (Un)fortunately I got a knee injury that required rehab and taking it easy for a while. During that time I got even more into computer science and tech as I had all this extra time all of a sudden (high school + 1-2 training sessions a day keeps you **busy**). Good thing this happened, really, because it was what built the foundation of my career. And I have a really good life because of it.

I stopped rowing at that point pretty much completely, but I remained active through gym and other workouts. I kept doing indoor rowing workouts and have always had a Concept2 rowing machine in my posession no matter where I lived. Later, in the past 5 years, I started two new sports: cycling and skiing. I transitioned to gravel cycling specifically over the past 3 years and added backcountry skiing. To keep my motivation up I signed up for the occasional gravel race and last year I did my first "ultra" (520km, 10k m vertical). And I loved it! For this year I plan that same ultra gravel race, [Seven Serpents: Quick Bite!](https://www.seven-serpents.com/quick-bite/) in May and the [World Rowing Master Regatta](https://wrmr2026.com) in September.

# Why

As you might be able to tell from the above background, I'm not your typical gym rat or weight-loss warrior, my focus is two-fold: develop strength and stamina for better life quality and do well in the above mentioned sports. I would put myself into the "amateur athlete" category. There's apps targeting this type of customer too but they often lack at least one major feature. And since we now live in the AI era, I looked at that as well and have not seen a good implementation of an AI coach targeting athletes just yet. I don't need fancy videos showing me how to perform an exercise. I need a very solid, knowledgeable and to-the-point kind of coach. I need a good and simple UX around seeing the training plan, performing workouts and syncing those to my Wahoo bike computer. The coach should remember a lot about the workouts and me. Oh and I made sure it has more of a to-the-point kind of vibe. No fluff.

# How

It's pretty simple actually. The frontend is a Next.js web app and an iOS native companion app. There's a Postgres database. That's it. The Next.js app handles everything from the frontend to the mobile API. Would I do it like this if I was not vibecoding it? Unlikely. I'd probably split things up a bit more, use multiple services, etc. But honestly, that would be over-engineering. Oh and also, I should say that while I do use the term "vibe coding" it's not really vibe coded. I go to great lengths to review the code but most importantly make sure the architecture and structure is sound.

I started by doing the work with Cursor. I like the integrated review capability, the editor and the fact that you can switch between multiple models. I used mainly Opus 4.5, some GPT-5.2 and the Cursor model, Composer 1, as well. Their model works surprisingly well for refactors and other simple(r) tasks, especially due to its speed. The iteration is super fast. However, what ended up happening as the project grew larger is that I did not really review all the agent changes in real-time. I would end up creating PRs, reviewing those and even then I'd help myself with agent PR reviews. The Cursor interface quickly became a hindrance (especially since it kept changing so fast). Then I briefly tried the Cursor CLI with the idea to keep the model-switching capability but I did not have a good time with that. It's so far behind the other good coding agent CLIs like OpenCode, Codex and Claude Code. I switched to using Claude Code as the main CLI with the occasional use of Codex and OpenCode as well. I like Claude Code's interactivity but for some specific tasks I prefer how Codex and OpenCode need a lot less hand-holding.

# What

This is how Claude describes the project:

> Fitter is an AI-powered workout coaching platform that combines a native iOS app with a Next.js web dashboard to deliver personalized fitness guidance. At its core is an
  intelligent AI coach that conducts onboarding interviews to understand your training background, goals, and physical profile—then provides
  contextual advice and can directly create or modify workouts on your behalf. The platform supports both cardio workouts (cycling, running, rowing, swimming) with structured
  intervals and training zones (Z1-Z5), and strength workouts with sets, reps, and exercise circuits. Users can plan their training via a visual calendar, sync workouts with Wahoo
  devices through OAuth integration, and receive weekly check-in notifications on iOS prompting reflection on recovery, fatigue, and motivation. The AI coach automatically takes
  notes during conversations—tracking feedback about workout difficulty, pain, or performance—building a memory that makes future advice more personalized. Built with Swift 6 and
  SwiftUI on mobile and TypeScript/React on web, Fitter bridges the gap between generic workout apps and having an actual coach who knows your history and adapts your training
  accordingly.

(Fitter is the working name I use for the project.)

![Coach chat](/images/2026-02-03/coach-chat.png)

# I like it because it's mine

I have a huge bias, obviously, but I think it has a unique angle on the AI fitness coach theme. It's not just another gym coach or workout manager, it's not an app to make you a better runner and it's also not an app to log and publish your workouts (keep using Strava for that, I do). But if your goal is to become stronger, fitter and do well in your sport(s) of choice, then it might be interesting to you as well.

Note: This is very much work in progress. I have been using it for months though and if you'd like to try it out let me know! Hit me up on [LinkedIn](https://www.linkedin.com/in/jernejstrasner/) or [X](https://x.com/JernejStrasner).


Building My AI Fitness Coach


Most of my career was iOS. I shipped apps, lived in Xcode, and got very used to the nice, tight feedback loop of building for one platform. At Facebook (now Meta), I expanded that into working with custom frameworks and tools outside the typical iOS development workflow. Later I moved into management and leadership work, which meant less hands-on coding day to day. Everything else (backend, frontend, databases, etc.) mostly happened on evenings and weekends as side projects. After I left Sentry, that weekend-and-evening hacking turned into full-time exploration.

As part of my exploration of LLMs and related infrastructure, I inevitably came across tool calling.

I’ve been building a workout manager and AI workout coach for myself ([fitness.antlab.ai](https://fitness.antlab.ai)). This gateway work grew out of my own need: I want models to take actions safely and predictably, not just generate text confidently. More on the AI coach another time. I think it’s a cool product in itself, and it’s gotten really good feedback from a few professional and amateur athlete friends who’ve been using it.

## Why build a gateway?

Tool calling is the part that makes LLMs useful beyond chat. Instead of guessing, the model can fetch data or run an action. The problem is that every provider implements “tool calling” with slightly different semantics, and wiring tools plus OAuth into every backend service becomes a mess quickly.

A gateway gave me three practical wins:

- Centralised configuration: define tools once, expose them to different models.
- Authentication and permissions: handle OAuth and per-tool scopes in one place.
- Consistency: normalise provider responses so callers do not need a pile of special cases.

My longer-term goal was more ambitious: manage context in the gateway and eventually stop depending on MCP servers. Even before that, the edge cases taught me where the spec ends and reality begins.

## Edge cases and implementation lessons

### 1. Resource metadata discovery is not consistent

Different MCP servers expose OAuth resource metadata in different places. What ended up working in practice was checking three locations, in this order:

1. The resource URL itself
2. `/.well-known/oauth-protected-resource` appended to the resource URL
3. `/.well-known/oauth-protected-resource` at the host level

This is one of those “I wish it was simpler” parts. You cannot assume everyone follows the same discovery endpoint.

### 2. `WWW-Authenticate` parsing is easy to botch

Some providers return a 401 with a `WWW-Authenticate` header that includes a `resource_metadata` URL. Parsing that header is fiddly, and it is very easy to accept something malformed (or reject something valid).

In my implementation there is still a TODO to add proper unit tests for this case, because it is exactly the kind of bug that only shows up in production.

### 3. Specs leave room, and providers take it

There are a few validation checks I still consider “must do”, but they take time:

- Supported endpoint auth methods
- Supported response types
- Supported grant types
- Refresh token support

The spec leaves room for interpretation, and providers use that freedom.

### 4. Discovery can fail even when you have credentials

One pragmatic change that improved UX: if discovery fails but the gateway has preconfigured OAuth credentials (some providers need this), it still populates client ID/secret so users can try authenticating. Some providers have flaky metadata endpoints, and letting the user proceed is better than a hard stop.

### 5. Token refresh needs a bit of slack

Tokens expiring “right now” is a common failure mode. I added a 30-second skew so the gateway refreshes slightly before expiry and avoids race conditions.

## Termination reasons: normalised vs native

Models report “why I stopped generating” in different ways. The gateway tracks two fields:

- `finishReason`: a normalised reason (for example `stop`, `tool_calls`, `length`)
- `nativeFinishReason`: the provider’s raw reason (for example `STOP` or `MAX_TOKENS`)

Here’s a simplified version of the check that decides whether to warn the caller about an upstream failure:

```go
func shouldNotifyUpstreamFailure(term upstreamTermination) bool {
    // Notify if nativeFinishReason indicates an error (non-empty and not "stop" or "max_tokens" - case insensitive)
    if term.nativeFinishReason != "" {
        native := strings.ToLower(term.nativeFinishReason)
        if native != "stop" && native != "max_tokens" && native != "function_call" && native != "tool_calls" {
            return true
        }
    }

    // Notify if finishReason indicates an error (not "stop", "tool_calls", or "length")
    if term.finishReason != "" {
        reason := strings.ToLower(term.finishReason)
        if reason != "stop" && reason != "tool_calls" && reason != "length" {
            return true
        }
    }
    return false
}
```

Two patterns that mattered:

- Gemini tends to use uppercase stop codes like `STOP` and `MAX_TOKENS`, while others use lowercase. Lowercasing avoids false alarms.
- Some APIs use `tool_calls`, others use `function_call`. In both cases it usually means “the model successfully decided to call a tool”, so I treat them the same.

Anything else (for example `content_filter`) triggers a warning, because it often means the response did not complete normally.

## How models differ in tool calling

The uncomfortable truth is that tool calling is learned behaviour. The API surface looks similar, but models make different choices about when to call tools, how strict they are about schemas, and how often they “give up” and answer in plain text.

### OpenAI (ChatGPT, GPT-4 class models)

OpenAI gives you explicit control over tool choice (auto, required, none). In practice, I’ve found behaviour varies a lot by model and prompt style. Some models will happily call a tool early, others need a very clear nudge, and the same toolset can behave differently across versions.

What helped me most was keeping tool names and parameter descriptions short, and not exposing too many tools at once.

### Anthropic (Claude)

Claude is generally comfortable mixing tool use with text, and it can chain tool calls naturally. It can also be more proactive than you expect, which is great when it is right and annoying when it is not. I ended up tuning tool descriptions more carefully to reduce “helpfulness” that turned into unnecessary tool calls.

### Google (Gemini)

Gemini has a lot of surface area and a lot of knobs, and when it works it can be excellent. The reliability was the harder part for me. I’ve seen runs where it calls tools correctly, then later in the day it answers in plain text instead of invoking the same tool, or it flips between behaviours across requests. That unpredictability is what pushed me to invest in stricter monitoring around termination reasons and tool-calling outcomes.

## Takeaways

Tool calling is not a pure spec implementation. It is model behaviour. If you want it to feel consistent to users, you have to do the boring work:

- Design tools for clarity. Short names, tight schemas, good descriptions.
- Do not expose everything at once. Too many tools lowers accuracy.
- Normalise and monitor. Treat unexpected termination reasons as real signals, not noise.
- Test across models. The same prompt and toolset will behave differently.

## Future work (and a note on the AI coach)

This gateway is still very much iterative. The next things I want to tackle:

- Better tool definitions, with examples that steer models toward correct calls.
- Lightweight JS tools so users can define simple functions without running a full MCP server.
- Automated regression tests that run prompts across models to catch behaviour drift.

And yes, the other project is still going: Fitter is a structured workout manager and AI coach. It lets athletes plan sessions, generate intervals, and sync workouts to devices. The coach side is where tool calling really earns its keep, but that’s a separate post.

## Conclusion

Tool calling is what turns LLMs into something you can build a product around, but it gets messy fast across providers. Finish reasons are inconsistent, tool-calling semantics vary, and reliability can swing from model to model. A gateway helps, but only if you treat standards as guidelines, normalise responses aggressively, and keep tightening your tool definitions based on real behaviour.


Building an LLM Gateway: Lessons from Tool Calling Across Models


I’ve been building products and tools for years — mostly iOS, lots of scripting, some backend, a lot of “make it work and ship it.” Recently I started poking into the AI space. Not just “call an LLM and magic happens,” but the *internals*. I want to understand enough to eventually build something serious on top of this tech.

One of the first things I kept hearing about in RAG (retrieval-augmented generation) setups was **chunking** and **embeddings** — and how they all tie into a **vector database**. I figured I’d document my beginner mental model.

## Step 1: Chunking
This part made sense quickly. You have long documents, but your embedding model (and later the LLM) has token limits. So you split the text into smaller **chunks**.

- Typical chunk size: ~500–800 tokens (2,000–3,200 characters in English)  
- Overlap: 10–20% so context flows between chunks  
- Too big = you get noisy, unfocused vectors. Too small = you lose context.

I now think of chunking as just a preprocessing step to keep everything bite-sized for the next phase.

## Step 2: Embeddings
Once you have chunks, you pass each one through an **embedding model**. This is where my “OK, but how?” curiosity kicked in.

The model outputs a **vector** — literally just an array of floats, like: [0.0213, -0.1342, 0.0876, …]

The magic is that similar meanings end up with vectors that are close together in this high-dimensional space. “Car” and “Automobile” will be neighbors. “Car” and “Banana” will be far apart.

It’s nothing like Levenshtein distance (string edits). Instead, embeddings are trained so semantic closeness translates to geometric closeness.

## Step 3: Storing in a Vector DB
Each chunk’s vector goes into a **row** in the vector database, alongside the original text and metadata. The DB builds an index so later we can search fast.

The vector DB doesn’t care about tokens or text anymore — it’s just arrays of numbers and math.

## Step 4: Retrieval
Here’s the clever bit:

1. User asks a question.  
2. You embed the question into its own vector (same model).  
3. The vector DB finds the **nearest neighbors** to that query vector.  
4. You fetch the corresponding chunks and feed them into your LLM for the final answer.

The naive way is to just scan all rows and pick the highest cosine similarity score (cosine similarity = dot product normalized for vector length). That works for small data, but with millions of vectors it’s too slow.

## Step 5: How Lookup is Actually Fast
Most vector DBs use an **approximate nearest neighbor** index — HNSW is the popular one. My mental model:

- Imagine all your points as cities on a map.  
- The DB builds a set of “highway” layers to jump across the space quickly, then “street” layers to refine locally.  
- When searching, it hops across highways, drops down to streets, and zeroes in on the closest points without visiting every “city.”

This gets you *almost* the same result as scanning everything, but way faster.

## Dot Product & Cosine Similarity (The Core Math)
The dot product is just multiplying each vector component, summing it up.  
Cosine similarity is the dot product divided by the product of vector lengths — which basically strips out magnitude and keeps only the **angle** between the vectors.

For embeddings, angle matters more than length, because we care about **direction in semantic space**, not absolute size.

## Final Mental Picture
- Chunk → embed → store in vector DB → embed query → find nearest vectors → send chunks to LLM.  
- Embeddings are just numbers.  
- Vector search is just “find points closest to this one” with clever shortcuts.  
- The rest is plumbing and product thinking.

I’m only scratching the surface, but already this feels a lot less like magic and more like a set of tools I can reason about.

Learning How Chunking, Embeddings, and Vector Databases Actually Work


As a (bored? topic for another post) startup founder currently between ventures, I've been methodically evaluating AI development tools to understand their practical applications and limitations. After extensive testing, particularly with Cursor and Claude Code, the results vary significantly across different technology stacks.

## The iOS Reality Check

My primary background is iOS development, where I've spent years building native Swift applications. AI tools in this domain remain frustratingly limited. The Apple ecosystem's complexity—from UIKit intricacies to SwiftUI's evolving patterns, Xcode's quirks, and platform-specific constraints—proves challenging for current AI models. While AI can generate basic Swift code, it struggles with the nuanced understanding required for production iOS development. It often writes outdated code, particularly when it comes to Swift 6 and concurrency. It can quickly iterate itself into a downward spirall of overcomplicated code which can be a complete mess. The feedback loops are slow, and the tools lack deep integration with Xcode workflows that iOS engineers rely on daily.

I've experimented with MCP implementations for Xcode, but the integration remains rudimentary compared to what's needed for serious iOS work. Claude often calls the tool wrong, tries a few times, and only then gets it right. This can be somewhat mitigated by having a well written CLAUDE.md file but it's not perfect.

## Web and Backend: A Different Story

The contrast with web development is stark. When working outside my primary domain—web frontend and backend development—AI tools prove remarkably effective. I recently built a complete [web version](https:///www.swimtemps.com) of my iOS app Swim using pretty much exclusively Claude Code. What would have taken weeks of research, boilerplate generation, and stack decisions was completed in a fraction of the time.

For Rust backend development, similar results emerge. The AI handles dependency management, error handling patterns, and API design with competence that approaches what I'd expect from a junior engineer familiar with the stack. However, due to Rust's intricacies it will still get things wrong. Armin Ronacher [recently wrote](https://lucumr.pocoo.org/2025/6/12/agentic-coding/) about LLMs being quite good at Go so I will give that a try next (and coincidentally, we chose Go as the primary backend stack at Specto for similar reasons why it works well with LLMs).

## Interaction Paradigm

The most significant shift isn't the code generation itself, but the interaction model. Rather than wrestling with command-line syntax or documentation, I can communicate with AI as I would with an engineer on my team. This conversational interface for system administration, project setup, and routine development tasks represents a meaningful productivity improvement. I generally run up to 3 terminal tabs with Claude in each one, either working on a different codebase/repository or even a different checkout of the same one.

I've been experimenting with bypassing permission prompts to increase iteration speed—a calculated risk that requires robust backup systems but dramatically improves the development flow. The AI maintains context across tasks, enabling complex multi-step operations that would typically require careful orchestration.

Claude Code is much easier to interact with in this new paradigm than Cursor. It lives in the terminal, focuses purely on the agentic aspect and doesn't even provide an editor. Cursor, as a vscode fork, is much more complicated and has a ton of options. Now ironically this post is being written in Cursor with a markdown preview plugin so there's still place for editors but I think Claude Code is on a better path.

## MCP and Automation

Beyond code generation, I've tested MCP implementations for browser automation with Puppeteer. The ability to describe user interface interactions in natural language and have them translated to reliable automation scripts opens interesting possibilities for testing and data collection workflows. This works pretty well most of the time but again, a well written CLAUDE.md file works wonders. As mentioned above, this works to some extent with Xcode as well but it's much more cumbersome. It will be interesting to follow the progress there. I've also tried using an MCP server for Railway (which is where I host most of my projects) and Vercel. For both I found it much easier to just get Claude to use the CLI tools for which I provided basic instructions in the respective CLAUDE.md files. The contents are basically the outputs of `vercel --help` and `railway --help`, condensed and with instructions to call the tool and command with `--help` to learn more when something doesn't work. Work wonders.

## Assessment

The current state suggests AI development tools are most valuable when working outside your core expertise. For an iOS engineer like myself, they excel at web development tasks I'd otherwise need to research extensively. The tools serve as force multipliers for exploring adjacent technologies rather than replacing deep domain knowledge (I'm convinced this will change super quickly though).

The interaction paradigm represents the more fundamental shift. We're moving toward intent-driven development where you describe desired outcomes rather than specifying implementation details. This abstraction works well for certain problem domains while remaining limited in others.

For now, AI development tools are powerful assistants rather than replacements, with effectiveness varying significantly by technology stack and problem complexity. It is starting to feel like you can start a company with less employees, particularly around building a website, blog, doing marketing, etc. The non-coding aspect is something I will explore more of next (I just connected Claude to my gmail inbox...).

**Disclaimer:** This post was written by me but proofread and formatted by Claude. Yet another way Claude helps me lately, especially as a non-native English speaker!


Jernej Štrasner

Blog

My heat pump's cloud app was stuck in 2012, so I reverse-engineered the protocol and cut it out

Building My AI Fitness Coach

Building an LLM Gateway: Lessons from Tool Calling Across Models

Learning How Chunking, Embeddings, and Vector Databases Actually Work

AI Development Tools: A Pragmatic Assessment