Sami Kiias
@samikiias.bsky.social
Head of Technologies @ Rascal AI | Empowering entrepreneurs through AI automation
AI is not always the right tool – here are 3 simple rules.
If a model doesn’t follow your instructions exactly, the problem isn’t always in the prompt. Sometimes it’s choosing the wrong tool for the job.
If a model doesn’t follow your instructions exactly, the problem isn’t always in the prompt. Sometimes it’s choosing the wrong tool for the job.
February 17, 2026 at 9:41 PM
AI is not always the right tool – here are 3 simple rules.
If a model doesn’t follow your instructions exactly, the problem isn’t always in the prompt. Sometimes it’s choosing the wrong tool for the job.
If a model doesn’t follow your instructions exactly, the problem isn’t always in the prompt. Sometimes it’s choosing the wrong tool for the job.
Prompt engineering" is starting to become an outdated term—now the name of the game is workflow engineering. We are moving from AI hype to a phase where simply chatting with a smart bot is no longer enough for productive work.
February 16, 2026 at 5:38 AM
Prompt engineering" is starting to become an outdated term—now the name of the game is workflow engineering. We are moving from AI hype to a phase where simply chatting with a smart bot is no longer enough for productive work.
Zapier was too expensive. Coding was too slow. I had to find something in between.
February 14, 2026 at 6:07 AM
Zapier was too expensive. Coding was too slow. I had to find something in between.
Right now the smartest developers and founders aren't just "chatting" with AI anymore. They're building pipelines. 🏗️
One signal has risen above everything else in technical communities lately: automation tools like n8n are evolving from simple data movers into intelligent agents. 🧵
One signal has risen above everything else in technical communities lately: automation tools like n8n are evolving from simple data movers into intelligent agents. 🧵
February 9, 2026 at 6:08 AM
Right now the smartest developers and founders aren't just "chatting" with AI anymore. They're building pipelines. 🏗️
One signal has risen above everything else in technical communities lately: automation tools like n8n are evolving from simple data movers into intelligent agents. 🧵
One signal has risen above everything else in technical communities lately: automation tools like n8n are evolving from simple data movers into intelligent agents. 🧵
What if AI's biggest disruption isn't code quality, but the maker's self-confidence?
This angle has been emerging surprisingly strongly in deeper developer community discussions. Instead of debating benchmarks, people are talking about something far more human. 🧵
This angle has been emerging surprisingly strongly in deeper developer community discussions. Instead of debating benchmarks, people are talking about something far more human. 🧵
February 8, 2026 at 1:08 PM
What if AI's biggest disruption isn't code quality, but the maker's self-confidence?
This angle has been emerging surprisingly strongly in deeper developer community discussions. Instead of debating benchmarks, people are talking about something far more human. 🧵
This angle has been emerging surprisingly strongly in deeper developer community discussions. Instead of debating benchmarks, people are talking about something far more human. 🧵
Stop guessing. Too many small businesses do marketing based on 'I feel like this might work.' I did it too. When I was a massage therapist, intuition worked — but in business, it’s like throwing darts with your eyes closed.
February 7, 2026 at 12:35 PM
Stop guessing. Too many small businesses do marketing based on 'I feel like this might work.' I did it too. When I was a massage therapist, intuition worked — but in business, it’s like throwing darts with your eyes closed.
"Sami, what AI should we get?" – That’s the wrong question. As a massage therapist, I learned that where clients feel pain is rarely the real cause. The same goes for technology.
February 7, 2026 at 12:34 PM
"Sami, what AI should we get?" – That’s the wrong question. As a massage therapist, I learned that where clients feel pain is rarely the real cause. The same goes for technology.
"Sami, what AI should we get?" – That’s the wrong question. As a massage therapist, I learned that where clients feel pain is rarely the real cause. The same goes for technology.
February 6, 2026 at 3:47 PM
"Sami, what AI should we get?" – That’s the wrong question. As a massage therapist, I learned that where clients feel pain is rarely the real cause. The same goes for technology.
Honestly? The first version of Rascal AI collapsed under its own weight. Airtable was chosen for scalability, but as user numbers grew, the server struggled like an old bicycle. Queries took forever, and I spent nights wondering how to survive without a million-dollar budget.
February 4, 2026 at 10:06 AM
Honestly? The first version of Rascal AI collapsed under its own weight. Airtable was chosen for scalability, but as user numbers grew, the server struggled like an old bicycle. Queries took forever, and I spent nights wondering how to survive without a million-dollar budget.
40,000 lines melted down to 3,000. Overnight. While I was asleep.
February 2, 2026 at 8:03 AM
40,000 lines melted down to 3,000. Overnight. While I was asleep.
I used to think building a startup meant polishing plans in a basement. Then I realized the real work begins when you throw your unfinished idea into the public.
January 30, 2026 at 12:35 AM
I used to think building a startup meant polishing plans in a basement. Then I realized the real work begins when you throw your unfinished idea into the public.
I used to be the masseur who thought time was money. Until I realized time is actually freedom – and technology can buy you more of it.
January 27, 2026 at 1:23 AM
I used to be the masseur who thought time was money. Until I realized time is actually freedom – and technology can buy you more of it.
Last year I tested over 50 AI tools. Most of them? Honestly, trash. But a few gems save me at least one full workday every week.
January 19, 2026 at 1:29 PM
Last year I tested over 50 AI tools. Most of them? Honestly, trash. But a few gems save me at least one full workday every week.
Many people think automation means blowing up all company processes at once. In reality, the biggest savings often hide in those boring routines done just because 'that’s how it’s always been done.' Sami, where should I start?
January 16, 2026 at 11:32 AM
Many people think automation means blowing up all company processes at once. In reality, the biggest savings often hide in those boring routines done just because 'that’s how it’s always been done.' Sami, where should I start?
The biggest misconception about AI? That you need to be some kind of tech wizard to use it. The truth is: people make this way too complicated.
January 14, 2026 at 1:39 AM
The biggest misconception about AI? That you need to be some kind of tech wizard to use it. The truth is: people make this way too complicated.
Imagine a bug report comes in, and an AI agent on a Raspberry Pi fixes it while you sleep. 😴🛠️
I just set up the infra: n8n triggers an SSH command to my Raspi, where Claude Code analyzes the repo and applies fixes. The age of autonomous coding is wild.
#AIagents #DevOps #n8n #Claude
I just set up the infra: n8n triggers an SSH command to my Raspi, where Claude Code analyzes the repo and applies fixes. The age of autonomous coding is wild.
#AIagents #DevOps #n8n #Claude
January 12, 2026 at 6:38 PM
Tässä on mun tapa tuottaa tekstiä, ilman että kosken näppäimistöön. Mä oon rehellisesti vähän kärsimätön kirjoittaja. Aivot käy ylikierroksilla, mut sormet laahaa perässä. Se on rasittavaa, ku ajatus on jo maalissa, mut kädet vasta hamuilee oikeita kirjaimia
January 12, 2026 at 12:37 PM
Tässä on mun tapa tuottaa tekstiä, ilman että kosken näppäimistöön. Mä oon rehellisesti vähän kärsimätön kirjoittaja. Aivot käy ylikierroksilla, mut sormet laahaa perässä. Se on rasittavaa, ku ajatus on jo maalissa, mut kädet vasta hamuilee oikeita kirjaimia
The AI hype is over. And that’s the best thing that’s happened to entrepreneurs in a while. Let’s be honest — we’ve had a bit too much sci‑fi dreaming and too little actual profit left on the line.
January 10, 2026 at 6:00 AM
The AI hype is over. And that’s the best thing that’s happened to entrepreneurs in a while. Let’s be honest — we’ve had a bit too much sci‑fi dreaming and too little actual profit left on the line.
"Make or n8n?" This is the most common question in automation. And most people approach it wrong. I hear it almost every week — for good reason.
January 7, 2026 at 6:00 AM
"Make or n8n?" This is the most common question in automation. And most people approach it wrong. I hear it almost every week — for good reason.
AI spitting nonsense? 3 easy ways to fix it fast.
Honestly, does it feel like AI only produces generic or confusing content? I know that feeling. You try ChatGPT and think: “Nice toy, but not useful for real work.”
Honestly, does it feel like AI only produces generic or confusing content? I know that feeling. You try ChatGPT and think: “Nice toy, but not useful for real work.”
January 6, 2026 at 3:37 AM
AI spitting nonsense? 3 easy ways to fix it fast.
Honestly, does it feel like AI only produces generic or confusing content? I know that feeling. You try ChatGPT and think: “Nice toy, but not useful for real work.”
Honestly, does it feel like AI only produces generic or confusing content? I know that feeling. You try ChatGPT and think: “Nice toy, but not useful for real work.”
"Learn to code" is dangerous advice for an entrepreneur in 2026. Many still think automation requires a nerd's brain and an engineer’s degree—that you must deeply understand ones and zeros.
January 5, 2026 at 12:06 PM
"Learn to code" is dangerous advice for an entrepreneur in 2026. Many still think automation requires a nerd's brain and an engineer’s degree—that you must deeply understand ones and zeros.
January 2026. Many people start the year with a massive “new life” project. Honestly? I don’t have the energy for that.
January 2, 2026 at 1:32 AM
January 2026. Many people start the year with a massive “new life” project. Honestly? I don’t have the energy for that.
Years ago, my main measure of success was how exhausted I was after work. If my brain wasn’t total mush by evening, I thought I hadn’t worked hard enough. That was a big mistake.
December 30, 2025 at 3:16 AM
Years ago, my main measure of success was how exhausted I was after work. If my brain wasn’t total mush by evening, I thought I hadn’t worked hard enough. That was a big mistake.
A year ago, I had a “real job” and constant busyness. Now I have a system and freedom.
December 26, 2025 at 2:32 PM
A year ago, I had a “real job” and constant busyness. Now I have a system and freedom.
A year ago, my Mondays started with the same ugly routine: scrolling through emails, moving data into the CRM, and juggling Excel sheets. It was “work,” but it drained all my creative energy right from the morning.
December 22, 2025 at 6:07 PM
A year ago, my Mondays started with the same ugly routine: scrolling through emails, moving data into the CRM, and juggling Excel sheets. It was “work,” but it drained all my creative energy right from the morning.