Hacker News

Article URL: https://arxiv.org/abs/2503.19058

Comments URL: https://news.ycombinator.com/item?id=43515023

Points: 1

# Comments: 0

Article URL: https://www.vice.com/en/article/garmin-ticked-off-smartwatch-users-with-its-new-subscription-fee/

Comments URL: https://news.ycombinator.com/item?id=43514999

Points: 7

# Comments: 0

Article URL: https://webkit.org/blog/16587/item-flow-part-1-a-new-unified-concept-for-layout/

Comments URL: https://news.ycombinator.com/item?id=43514997

Points: 2

# Comments: 0

I often create presentations where elements need to move (animate) from one slide to the next. This is great for explaining algorithms, workflows, or anything that benefits from dynamic visuals.

I used to rely on Figma prototypes with smart animations for this, but I wanted a way to do it in Excalidraw—maybe even more conveniently. So I made an Excalidraw fork which helps you present "frames" and interpolates animations between them.

Video demo (detecting cycles in a graph): https://github.com/user-attachments/assets/62033fca-03ca-489...

Instructions, tips, and current limitations here: https://github.com/excalidraw-smart-presentation/excalidraw-...

Try it here: https://excalidraw-smart-presentation.github.io/

Personal plug: I'm also currently looking for a full-time job. If you know of any opportunities or can refer me, I'd really appreciate it.

Comments URL: https://news.ycombinator.com/item?id=43514976

Points: 1

# Comments: 0

Hey HN,

I’ve always found it frustrating to discover good content creation tools—there are thousands out there, but:

Most lists are bloated with low-quality or outdated tools.

AI hype is everywhere, making it hard to separate useful tools from gimmicks.

Pricing isn’t always transparent, and affordable/free tools are buried.

The real challenge? Finding the right tools for specific needs—like: Turning podcasts into YouTube Shorts effortlessly. Writing high-engagement LinkedIn posts with AI assistance. Creating viral TikToks with easy-to-use video editors. Generating email copy optimized for conversions.

So I built ContentCreation.tools—a curated directory of content tools, with: No-bullsh*t lists—only high-quality tools. Budget filters—find free/affordable options easily. Curated tool stacks—get the right set of tools for specific tasks.

Would love to hear your thoughts! What are your biggest frustrations when searching for the right tools? How do you currently find them?

Comments URL: https://news.ycombinator.com/item?id=43514974

Points: 1

# Comments: 0

Article URL: https://paulgraham.com/do.html

Comments URL: https://news.ycombinator.com/item?id=43514969

Points: 1

# Comments: 0

Article URL: https://twitter.com/hardmaru/status/1684940023974793216

Comments URL: https://news.ycombinator.com/item?id=43514966

Points: 1

# Comments: 0

Article URL: https://electrek.co/2025/03/28/hyundai-evs-smarter-more-efficient-new-pleos-brand/

Comments URL: https://news.ycombinator.com/item?id=43514927

Points: 1

# Comments: 1

My brother and I have been working on Morphik: an open-source multimodal database. (Github: https://github.com/morphik-org/morphik-core)

After experimenting with various AI models, we realized that they were particularly bad at answering questions which required retrieving over images and other multimodal data.

That is, if I uploaded a 10-20 page PDF to ChatGPT, and ask it to get me a result from a particular diagram in the PDF, it would fail and hallucinate instead. I faced the same issue with Claude (not with Gemini, though).

Turns out, the issue was with how these systems ingest documents. It seems like both Claude and GPT embed larger PDFs by parsing them into text, and then adding the entire thing to the context of the chat. While this works for text-heavy documents, it fails for queries/documents relating to diagrams, graphs, or infographics.

Something that can help solve this is directly embedding the document as a list of images, and performing retrieval over that - getting the closest images to the query, and feeding the LLM exactly those images. This helps reduce the amount of tokens an LLM consumes while also increasing the visual reasoning ability of the model.

We've implemented a one-line solution that does exactly this with Morphik. You can check out the specifics in the attached blog, or get started with it through our quick start guide: https://docs.morphik.ai/getting-started

Here is an example ingestion pathway:

``` from databridge import DataBridge

db = DataBridge()

db.ingest_file("report_with_images_and_graphs.pdf", use_colpali=True) ```

And here is an example query pathway:

``` db.query("At what time-step did we see the highest GDP growth rate?", use_colpali=True) ```

Would love to hear your thoughts!

Comments URL: https://news.ycombinator.com/item?id=43514922

Points: 1

# Comments: 0

Article URL: https://en.wikipedia.org/wiki/Time_formatting_and_storage_bugs

Comments URL: https://news.ycombinator.com/item?id=43514920

Points: 1

# Comments: 0

Article URL: https://koto.dev/

Comments URL: https://news.ycombinator.com/item?id=43514915

Points: 1

# Comments: 0

Article URL: https://www.nature.com/articles/s41587-025-02584-1

Comments URL: https://news.ycombinator.com/item?id=43514908

Points: 1

# Comments: 0

Hey everyone! I recently built this little side project that takes any document you upload and turns it into practice quizzes using LLMs to generate the questions: https://www.cuiz-ai.com

Being a backend engineer, I have always struggled to finish my side projects due to my awful frontend skills. Now, with the progress of AI coding tools, I had no more excuses. 100% of the frontend, landing page and even the logo was made using a combination of Cursor, ChatGPT and Claude, but it's always been under my supervision and never blindly accepted the proposed changes. The backend and infra was handled by myself.

The tool itself is pretty straightforward: upload any PDF/DOC, get quiz questions generated by an LLM, export them if you want, and come back to your previous quizzes anytime. I got this idea when I was studying from a pdf for my UK citizenship. I know there are similar tools around, but it was a good opportunity for me to build something end to end!

I'd love to know what you think or what you might use this for. Any feedback is super appreciated!

Comments URL: https://news.ycombinator.com/item?id=43514898

Points: 3

# Comments: 0

Article URL: https://www.quantamagazine.org/the-high-cost-of-quantum-randomness-is-dropping-20250328/

Comments URL: https://news.ycombinator.com/item?id=43514816

Points: 1

# Comments: 0

Article URL: https://www.riscosopen.org/news/articles/2025/03/28/moonshots-initiative-to-secure-the-future-of-the-os

Comments URL: https://news.ycombinator.com/item?id=43514803

Points: 3

# Comments: 0

Article URL: https://phonepe.github.io/bonsai/index.html

Comments URL: https://news.ycombinator.com/item?id=43514790

Points: 3

# Comments: 0

Article URL: http://stephantul.github.io/python/typing/2025/03/28/overload/

Comments URL: https://news.ycombinator.com/item?id=43514786

Points: 1

# Comments: 1

Article URL: https://www.perfriend.com

Comments URL: https://news.ycombinator.com/item?id=43514757

Points: 1

# Comments: 0

Article URL: https://christinapagel.substack.com/p/how-to-fight-back-charting-opposition

Comments URL: https://news.ycombinator.com/item?id=43514753

Points: 3

# Comments: 0

I keep seeing those tweets and posts where users ask ChatGPT or a similar LLM to describe them etc... and it always answers positive cool stuff which reinforces what the user wants to hear.

If you also try to ask it about a certain topic or yourself, it will always be positive and agree with your opinion. I feel there is a lot of confirmation bias at play.

Comments URL: https://news.ycombinator.com/item?id=43514752

Points: 1

# Comments: 1