Hacker News new | past | comments | ask | show | jobs | submit | buss_jan's comments login

I find it funny that is message resurfaces on the front page once or twice a year for at least 10 years now. Product quality is often not the main argument advanced when deciding on a tech stack, only indirectly. Barring any special technical requirements, in the beginning what matters is: - Can we build quickly without making a massive mess? - Will we find enough of the right people who can and want to work with this stack? - Will this tech stack continue to serve us in the future?

Imagine it's 2014 and you're deciding between two hot new framework ember and react, this is not just a question about what is hot or shiny and new.


Great idea! I would totally swipe through. One small recommendation, consider having the title and a short description written by an LLM making it as dumb and sensational as possible, for a true tiktok like experience for instance the this title and abstract could be transformed into the following

Instead of:

"Benchmarking Multimodal RAG through a Chart-based Document Question-Answering Generation Framework

Multimodal Retrieval-Augmented Generation (MRAG) enhances reasoning capabilities by integrating external knowledge. However, existing benchmarks primarily focus on simple image-text interactions, overlooking complex visual formats like charts that are prevalent in real-world applications. In this work, we introduce a novel task, Chart-based MRAG, to address this limitation. To semi-automatically generate high-quality evaluation samples, we propose CHARt-based document question-answering GEneration (CHARGE), a framework that produces evaluation data through structured keypoint extraction, crossmodal verification, and keypoint-based generation. By combining CHARGE with expert validation, we construct Chart-MRAG Bench, a comprehensive benchmark for chart-based MRAG evaluation, featuring 4,738 question-answering pairs across 8 domains from real-world documents. Our evaluation reveals three critical limitations in current approaches: (1) unified multimodal embedding retrieval methods struggles in chart-based scenarios, (2) even with ground-truth retrieval, state-of-the-art MLLMs achieve only 58.19% Correctness and 73.87% Coverage scores, and (3) MLLMs demonstrate consistent text-over-visual modality bias during Chart-based MRAG reasoning. The CHARGE and Chart-MRAG Bench are released at https://github.com/Nomothings/CHARGE.git."

Give me: " "AI Fails Hard at Reading Charts: New Study Exposes Shocking Weaknesses!"

A groundbreaking study reveals that even the smartest AI models struggle with charts, scoring just 58% accuracy—proving your brain might still be better than AI at decoding data! "


Interesting idea. Is there a way to view the entire directory of tools by category, I would need to know what the search space is here. For examples I searched for "Webinar software that integrates well with Hubspot" and it only brought up one webinar tool with no Hubspot specific integration and two CRMs. Obviously not what I was looking for. I like how simple the interface is otherwise.

Very real problem. We have started to worry about how to maintain the many cookbook samples and tutorials, we wrote/are planning on writing. Going over every user facing artefact after a code change is a non-starter, so you opt for versioning these things as they'll necessary become out of date. Very excited to give this a spin soon.

Even OpenAI's cookbook is out of date in many places. This is a really difficult and common problem!

Yeah, even with lots of full-time resources, companies' documentation regularly fall out of date. It's particularly ubiquitous in AI just because of how fast the general space is changing.

Can AI really accelerate product development? I built a short video generator in one day using AI and CE.SDK to find out.


Mhh, if I have evolved to a point where my experienced feelings and ideas are of such complexity that animal cues no longer suffice to broadcast them to the group self-awareness might have involved in tandem with the more complex nogging and not as a downstream feature.


This looks great! What are your plans around community and content, I feel like Pitchfork could use some competition. Best of success!


Thanks for the comment!

We are working through what exactly that will look like at the moment.

First thing we think we want to add is comments to tracks.

Following this I want artists to meet other artists and to start sharing their fan base. Some kind of ‘connect’ feature that allows artists to chat and shows they support each other. Also if an artist shares a track they get mentioned on the home feed in some way.

Finally we want to add a full forum type thing like indie hackers as well as a blog for artist tips and advice.

Little by little for now though!


This is great. Chapeau. It's a little strange to select the muscle group one wants to target, what about a few primary and secondary dimensions such as explosiveness, endurance, strength, fat loss or aesthetics. Some way to more broadly describe what one is trying to accomplish.


The project is GPL licensed, if, however, you cannot use GPL licensed software do get in touch with support@img.ly and we'll make the package available to you under a different license.


Thanks for your interest, currently not but keyframing is on the roadmap for next year!


Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: