sdpy

There is a slide about Narwhals and Ibis here (a recording of a talk from PyData Berlin 2025): https://youtu.be/qOJiQtqI348?t=2496 TL;DR: It doesn't have to be Narwhals vs. Ibis because Narwhals supports Ibis.

Update: Here's a benchmark as well, but it may be outdated: https://www.kaggle.com/code/marcogorelli/narwhals-vs-ibis-ov...

Just to add, people in my network have been talking about polars (as an alternative to pandas) and other dataframe libraries. They're much easier to use now thanks to the Narwhals compatibility layer (for example, Narwhals was recently added as a dependency to scikit-learn).

I'm using Notion Web Clipper (as a replacement for Evernote Web Clipper) to save promising links for future reference. Then I close these tabs.

> Many agriculture products are prohibited entry into the United States from certain countries because they may carry plant pests and foreign animal diseases.

> Prohibited or restricted items may include meats, fresh fruits and vegetables, plants, seeds, soil and products made from animal or plant materials.

https://www.cbp.gov/travel/international-visitors/agricultur...

I've received it from ycoommbinator/-co

A 2024 podcast with Bending Spoons about their acquisitions approach and company culture: https://newsletter.pragmaticengineer.com/p/twisting-the-rule...

Location: Europe

Remote: Yes

Willing to relocate: Yes

Technologies: Python (scikit-learn, pandas, PyTorch, transformers, networkx, etc.). Open to working in other languages, for example, Rust, C#, R. Ability to rapidly learn and apply new technologies, methods, and research techniques.

Résumé/CV: https://drive.google.com/file/d/1obU2OOHXcZe0b8aYUDbRsrZkCZi...

Email: available in my CV

Summary:

- R&D software engineer with 10+ years of experience in data science, NLP, and AI/ML. Bachelor's and Master's degrees in Computer Science. Experience in R&D projects in academia and industry.

- Developed proof-of-concepts for startups. Built and improved NLP features for a SaaS product with over 1.5M users. Co-founded a startup that won a Demo Day in a startup accelerator in Italy.

- Comfortable with remote work and Agile/Scrum workflows.

  Location: Italy, Europe

  Remote: Yes

  Willing to relocate: Yes

  Technologies: Python (scikit-learn, pandas, PyTorch, transformers, networkx, etc.). Willing to learn whatever is needed for the job. Open to working in other languages, for example, Rust, C#, R. Machine learning, natural language processing (NLP), complex network analysis, data visualization (seaborn, matplotlib, Gephi, ggplot2). 

  Résumé/CV: https://drive.google.com/file/d/1vVFhqiVQNi6Cby-klqNb7uBS3FIvgF1h/view?usp=sharing

  Email: available in my CV

10+ years of experience in programming, data science, NLP, and AI/ML. Built and improved NLP features for a SaaS product with over 1.5M users. Developed proof-of-concepts for startups. Co-founded a startup that won Demo Day at a startup accelerator in Italy. Analyzed data and trained ML models. Maintained legacy code and fixed bugs. Open to any kind of backend work, not only ML/NLP, and can learn any required technologies. Comfortable with remote work and Agile/Scrum workflows.

That would be a game-changer for so many people. We're already getting close, though. Those CGMs like Dexcom don't actually test blood; they check interstitial fluid. There's a 5-15 minute delay, but still way better than finger pricks all day.

Non-invasive monitoring is still tricky. There’ve been some interesting attempts—like GlucoWatch back in the early 2000s, which used mild electrical currents, but caused skin irritation and never really caught on. Others have tried optical monitoring, radio waves, ultrasound, and even heat-based sensors. Feels like we’re getting closer, just not quite there yet.

There is also an open-source project Nightscout (CGM in the Cloud) which supports Android and iOS: https://github.com/nightscout

A similar idea appears in a more recent short story, How It Unfolds (2023) by James S.A. Corey. The premise is using a technology called “slow light,” which can clone people and objects using “enriched light.” The National Space Agency scans a group of 200 people (not only their physical forms, but also their consciousness, memories, and feelings) and transmits several thousand copies of this data package across the galaxy. The hope is that, on arrival, each package can "unfold" into a fully reconstructed version of the original team and habitat on some distant alien world.

The Prompt Report: A Systematic Survey of Prompting Techniques https://arxiv.org/abs/2406.06608

Some users chose Authy precisely because of the desktop app. Phones can be lost / stolen / unresponsive / out of battery, and it was nice to have a second option for 2FA.

------

A list of alternatives they provide:

> Authy Desktop Alternatives > If you are unable to use the Authy mobile apps, we recommend you look in to alternative desktop products such as the following suggestions:

> https://authenticator.cc/ > https://steptwo.app/ > https://secrets.app/ > https://keepassxc.org/ > https://support.1password.com/one-time-passwords/

> Please be aware that Twilio Authy does not have any affiliation with the companies creating these apps, so be mindful of potential risks and exercise your own judgment when selecting the app.

> Note: The Authy app lacks an export feature, therefore you have to re-enable the 2FA method and re-add the token in your new application with the steps here.

FAQ: https://community.upwork.com/t5/FAQ-Suspended-Operations-in/...

> But if you don't turn your TV on at all, I have to ask - where do you watch Netflix et al?

I watch video content on my 15.6" laptop with a retina display. The picture is subjectively larger, brighter, and sharper than on a TV because it's closer to my eyes. It's like in a movie theater but with lots of privacy. So I prefer it this way. I would turn on a TV only if I'd want to watch something with another person, I guess.

Relevant Twitter thread: https://twitter.com/billmarczak/status/1416817352549625858

Maybe this one: https://news.ycombinator.com/item?id=27876165 > MVT – forensic tool to look for infections on smartphones (github.com/mvt-project)

The data can be downloaded at https://accounts.moves-app.com/export?download=true

But, apparently, Move-O-Scope is not able to download any new data: > Error: he API package 'xmpp' or call 'CreateChannel()' was not found.

I am wondering if any of the other connected apps can still be used to display the data.

The original study: http://www.thoracic.org/about/newsroom/press-releases/resour...

There was a write-up about this paper in MIT Technology Review: https://www.technologyreview.com/s/610395/if-youre-so-smart-...