
Technology Advisor & Business Engineer 🇨🇭 Practical AI & Data Science insights for real work — not just demos. Zürich | #DataScience #AI #ML
Zürich CH
Joined April 2011
- Tweets53.1K
- Following2.7K
- Followers3.3K
- Likes8.9K
Pinned Tweet
64% of institutions see big potential in blockchain, 8% actually offer tokenized real estate. My take on the gap.
56
GPT-6.2 Astra and Sol: A Hands-On Performance Review via/AnalyticsVidhya #DataScience #Data #Analytics #AI #ML #India ift.tt/aSyI5WO
1
40
Gemini 4 Neon Review: Benchmarks, Pricing, and Performance via/AnalyticsVidhya #DataScience #Data #Analytics #AI #ML #India ift.tt/RsXMzTo
31
Thoughts on Verification Engineering in the Agentic SDLC via/startdataengineering.com
#data #engineering #AI #ML
ift.tt/oVsg6Mk
47
How to Use Claude for Stock Research, Portfolio Reviews and Trades via/AnalyticsVidhya #DataScience #Data #Analytics #AI #ML #India ift.tt/wuM0jo7
29
How to Make Product Launch Videos with Claude Opus 5.5 via/AnalyticsVidhya #DataScience #Data #Analytics #AI #ML #India ift.tt/IsTmdjx
36
Top 10 GitHub Repositories Trending in October 2026 via/AnalyticsVidhya #DataScience #Data #Analytics #AI #ML #India ift.tt/PWVf5aR
45
🌧Light Rain tomorrow in #Zürich! Don‘t forget to bring your umbrella ☔️
64
🌧AM Light Rain tomorrow in #Zürich! Don‘t forget to bring your umbrella ☔️
59
🌧Light Rain tomorrow in #Zürich! Don‘t forget to bring your umbrella ☔️
61
claudîus retweeted
Someone just dropped 1.4 million pixel art sprites on Hugging Face 😱👾
free to use, captioned, 16×16
huggingface.co/datasets/Anga…
Fast scraping is easy. Complete scraping isn't.
In this test, 7,225 pages were scraped in under 3 minutes. The full job took 22 minutes, and 62 of 10,000 pages were missing.
Count what actually arrived before you trust the dataset.
via @kdnuggets kdnuggets.com/web-scraping-7…
54
Time series in Python means a lot of libraries.
Good overview of 7: statsmodels for interpretable ARIMA, pmdarima for quick baselines, plus ML, deep learning and anomaly detection.
For price data, start with a model you can explain.
via @kdnuggets kdnuggets.com/top-7-python-l…
41
A capable AI agent on your own machine, no cloud model needed.
Hermes Agent (MIT license) with Ollama: local models, persistent memory, MCP tools and sub-agents.
Local is a good start for sensitive data. Pair it with a sandbox.
via @kdnuggets kdnuggets.com/what-i-did-wit…
3
3
123
claudîus retweeted
ChatGPT has now a big problem.
Researchers at Oxford exposed a massive threat to large language models.”
It’s called "Model Collapse."
And it explains why the internet is about to become a toxic waste dump for Large Language Models.
Right now, AI companies are scraping the entire web to train their next-generation models.
But there’s a massive problem.
The internet is no longer just human data. It is rapidly filling up with AI-generated text, images, and code.
What happens when an AI trains on data generated by another AI?
A new paper in Nature just gave us the answer. And it is terrifying.
When AI models are recursively trained on AI-generated data, they suffer from irreversible brain damage.
The researchers found that the model starts forgetting the "tails" of the original data distribution. The rare, weird, unique, and complex human nuances completely disappear.
It's like making a photocopy of a photocopy.
Every generation loses resolution. The output gets flatter, blander, and more generic, until the model completely collapses in on itself.
For years, we thought the biggest bottleneck for AI was compute, getting enough H100 GPUs from Nvidia.
This paper proves the real bottleneck is us.
If we keep flooding the internet with synthetic AI slop, we will poison the very data pool required to build smarter models.
The ultimate irony? We thought AI was going to make human creators obsolete.
Instead, this research proves that verified, genuine human data is about to become the most valuable commodity in the tech industry.





