661–670 of 1001 posts
I'm working on scoping out specific problems facing "software as a medical device" (SaMD) companies. In particular issues around being able to release software at a reasonable cadence. I've been a CTO in this space for a couple of years and I am now consulting with other firms around the intersection of tech and regulatory. It's a tight-rope walk of ensuring that all testing (software and non-software testing) and ev…
In my 50k population town in Minnesota, I notice a lot of headlights, taillights, and turn signals are out. Our roads (outside of the Twin Cities) are often bumpy which causes vibration in the vehicles, leading to many bulbs failing. I drive a 21 year old Saab, and in my 2 years of owning it, I have replaced every single bulb in the exterior of the vehicle except a turn signal or two. I decided to create a mobile ser…
Unless you are doing this purely as a public service, $10 seems VERY low. Many independent mechanics in my area charge a minimum of 1 hour of work for anything they do, and bill at $90/hr and up. And they won't come to you for that price. (Nor are mechanics a highly paid profession in general.)
I really do wish you all the best, maybe I'm just not seeing the whole picture.
Edit: Something else to think about: when I replace bulbs in my cars, I always do it in pairs. Reason being that if one side burned out, the other is likely not far behind, especially headlight bulbs. The headlights also have another quirk: if you replace one and not the other, you usually end up with the new one being much brighter than the old one.
Just doing plain OCR doesn't really work because the notes in the margin and the footnotes get mingled with the text, which results in gibberish.
But, when sent to Google Vision API, each page results in a json file that has an object for each word and the four coordinates of its bounding box.
That json file is pretty big (around 1.5 Mo when pretty printed, or 500 k with no indents or line breaks) but it can then be fed to Gemini, taking advantage of its large context window.
Gemini is pretty good at identifying each section of the page (headers, main text, margin comments, footnotes) but it takes a looong time to respond (2-5 minutes per page).
So another approach is to ask Gemini to write a python script to analyze the json result and group sections depending of the coordinates of each word, and then run that script against the json output by the OCR phase.
But it's quite difficult to have a script that works for any page; comments in the margin are always in the margin so that's pretty easy, but footnotes can start at any height of the page (some pages contain only footnotes running from previous pages) and Gemini likes to be pretty specific, giving hard 'y' coordinates for where footnotes should start, which obviously only works for the one page it's working on.
I'm iterating and making some progress but I feel like I miss a big breakthrough and it all should be simpler than it currently is. Information about OCR is pretty scarce online. Any pointer is welcome!
I've been obsessed with developing ways to make it easier to handle tab overload in the browser without requiring any sort of active "tab management". I have a working extension that replaces the "new tab" page with a clean view of all open tabs, along with simple ways to search and select which tab to switch to, including search over bookmarks and history. There are also some simple tools to allow for creating and r…
Some gripes:
I wish I could just use it standalone without it replacing the "new tab" page.
I wish I could view the full title of a tab, maybe on hover.
I only want to use it to manage tabs, not history or bookmarks, so I wish I didn't have to grant those permissions.
The main motivation for this is that I usually juggle many side projects at the same time, and since none of them are in a finished state, I don't really feel like putting them out in the open on GitHub or Codeberg or something else, and would really like to keep all of this experimentation private. I know GitHub provides free private repositories, but I really dislike the interface and I feel like most of the features offered are not for me.
I tried running a private Gitea instance, but I did not like having to manage a second software to be able to run CI pipelines on my repositories (like Drone). I also really dislike GitHub Actions so I was kind of disappointed that Gitea's integrated CI is based on it.
I also tried running a Gitlab instance (which is currently what I use) as I much prefer GitLab CI over GH Actions, but the whole package is so much of a resource hog and tries so hard to be "more than GitHub" that I feel it isn't for me either.
So, in the end, I decided that I would build my own platform, with a simple scope in mind: be a nice Git repository hosting solution for single hackers or very small teams. Along with the smaller scope, I added a few restrictions: no JavaScript unless it substantially improves QoL (and even then, never require JS for the website to work), no external components required (apart from OpenSSH and Git).
So far I have a very minimal SSH handler implementation that allows push/pulling repositories from the server through authenticated SSH (public key), and what's starting to become a nice web interface to manage and read all of that. It's almost at the point where I will soon be able to dogfood the platform by hosting itself there (as a backup remote, not the primary yet).
The tech stack is kept relatively simple on purpose (I really dislike "modern" front-end development):
* Rust + Axum
* SQLite
* Bootstrap with SASS to customize it
* libgit2
Everything is server-side rendered as HTML before being served to the client, using Minijinja (reimplementation of Jinja2 in/for Rust by the same author).
Anyway, not yet ready to share any links or anything but I am still excited about the project taking shape after weeks/months of sporadic work on it! I'm building this not only because I feel a need for it, but also to learn more about Rust and Git internals, and I feel like I've learned a ton already!
I'm experimenting with an AI assisted world building / story telling tool: https://youtu.be/OGkSI3VfxRU . The idea is to do kind of what Cursor is doing for programming; accelerate the creative process (rather than replacing it).
By the way you should get interested in "narratology" at large. There are benefits to driving the LLM generation with an explicit representation of the narrative structure. For instance if the bad guy turns out to be the good guy, then, narrative logic demands that the good guy was the bad guy all along. As a consequence, appearance and deceit should be keystones values of the hero's arch. There are many interdisciplinary approaches to this domain, some more naturalist/empirical such as trope databses, some theoretical like 70s semiotics while some other framework approach the problem from the perspective of cognition and entropy.
Even though I haven't really started coding anything, this is a very interesting topic to think about because of the technical challenges it raises as well as the scope of the project. Any papers exposing new LLM use and prompting techniques is of potential interest to me now.
It is a wide space
Now, there is a lot of CMS software out there. Some of the better ones are paid products.
What I'm hoping to eventually accomplish is easy local creation of a website (content and themes) and after that easy one click deployment to a cheap hosting provider. Alternatively just copying a local folder to your own vps/server with the CMS should be enough.
My dream outcome would be a CMS that is a one-stop solution for most types of websites (blogs, company sites, shops, ...). To hopefully contribute to making people stop using facebook, twitter, other centralized and eventually login seeking services for hosting content people would like to read.
For this, a free/cheap one click hosting solution after locally creating and previewing a site would be necessary.
PHP is still pretty widely used here because cheap web hosting package support that. I like PHP, but for open source projects I prefer Go because of the maintainability and fun of writing it.
I'm working on scoping out specific problems facing "software as a medical device" (SaMD) companies. In particular issues around being able to release software at a reasonable cadence. I've been a CTO in this space for a couple of years and I am now consulting with other firms around the intersection of tech and regulatory. It's a tight-rope walk of ensuring that all testing (software and non-software testing) and ev…
Now consulting? But yes I’d love to connect.