Building until I can surf every day 🏄 From Ecom guy to Saas Builder | Founder of https://xitter.cf/t.co/LpheJOSPZg Powered by sports and vibe coding
Joined August 2023
- Tweets1.2K
- Following374
- Followers228
- Likes1.6K
Built an Airtable CRM for a real estate agent with Claude.
Then asked it for a guide: 12 video tutorials, almost 35 minutes, with voice and subtitles. Plus a Notion guide with quizzes. For free and impressive quality
All from a few prompts. I love AI.
Spending weeks figuring out invoices, contracts and paperwork before you have your first client is a waste of time.
I'd rather start and solve problems as they come than plan for every possible obstacle. That's the difference between thinking and doing.
My medical history shouldn't depend on which country I live in.
I'm building a private vault for my blood tests, scans and reports.
The goal: find any document with AI and compare results year after year, wherever I am.
Worth doing. Keep control of your health data guys.
When you're researching a new business idea, what do you like doing the most?
1) Studying competitors
2) Researching products
3) Talking to customers
4) Spotting market trends
For me it's studying competitors. I really enjoy seeing what they've built, how they price it and what their customers say.
I'm building an open source tool but finding a good name is so hard. All the good names are already taken on GitHub.
Claude is getting pretty good at static ad creatives.
I gave it a competitor ad as a reference, and it recreated the layout by editing the elements, not generating a whole new image.
Honestly impressive. It costs me less than Nano Banana or GPT Image, with fewer retries and no huge prompt.
I studied my market with Claude tonight. Ouch, that was hard.
I asked it to audit my business like a McKinsey consultant. 50 questions + follow-up questions based on my answers (my brain was fucked by the end).
It questioned me about my customers until I saw what I knew, what I guessed, and where I contradicted myself.
It made me doubt a lot, but showed me what to research.
It's a funny exercise to see if you really know your market and your customers.
I upgraded my SEO agent for almost nothing today.
SerpApi tracks Google AI Mode citations. OpenRouter asks Perplexity and other AIs. Ahrefs Webmaster Tools checks issues and backlinks.
Good tools, models and process. Let's see if better analysis improves the results.
It feels like 2022 again. ChatGPT came out and suddenly everyone was building one.
Jev came out last week, and now Fastino has its own decision model. Feels a bit nostalgic watching the same race start again, but this time for models that make decisions.
Introducing GLiNER2.5-Decide, our new 340M parameter open weight, encoder-based decision model.
GLiNER2.5-Decide is built for fast, deterministic classification. The model evaluates a set of user-defined typed questions and rules, and jointly decodes their answers, returning structured decisions with probability distributions and confidence scores.
We evaluated the model’s performance on Fast Decisions, an unseen, internally generated classification suite based on 17 datasets testing real-world use cases across routing, triage, classification, sentiment, and content understanding.
Measured against similar decision models, GLiNER2.5-Decide leads in 9 of the 17 datasets, achieving the highest average score:
- GLiNER2.5-Decide: 60.1%
- SemIf: 56.4%
- JevK5: 57.5%
- Laya: 46.6%
This performance makes the model a strong fit for use cases like tool calling, model routing, browser and computer use, and LLM-as-a-judge.
GLiNER2.5-Decide’s lightweight encoder architecture makes it easy to fine-tune the model for specific tasks, while being efficient enough to run locally on consumer-grade CPUs or in air-gapped environments, giving users greater control over where their data goes and where the model runs.
To make building and experimenting with GLiNER2.5-Decide as easy as possible, we're also offering hosted inference. You can now use our API to run inference and fine-tune GLiNER models on specialized tasks right inside your own coding agent: agent.fastino.ai
As with previous models, we’re also releasing the model weights on @huggingface under the Apache 2.0 license: huggingface.co/fastino/GLiNE…
