SEEKING WORK | Remote US - Colorado based | QA/Test
QAComet is a one-man QA As A Service agency providing unlimited QA requests for $5k/month. There is a $50 14 day trial for the first three customers who contact me through HN. My services include
- E2E/Integration/Unit testing
- QA/Usability reviews
- Bug finding/reporting
You can see a case study of some past work I've done with Automatisch.io helping build their E2E test suite and other QA work here
This is a neat product, and I plan on trying out some of the recommendations for sunscreen.
During my journey using the app there were a few things I noticed
1) It seems like the intermediate page is generating text from the LLM as well, which makes the whole process quite slow on my machine. It took maybe 10 seconds before the loader finished displaying the text. If I try and perform the same query again on the same browser, the results are somewhat quicker, maybe 700-800ms of wait time, but this still seems too slow. Once I ran the query five or so times, it was as quick as the demo queries on the front page.
2) Consistent results: If I use the same query on separate browsers, I'm given different products as the "Top Recommended Product", which seems odd. I know LLMs are stochastic, but the feed starting with the "Top Recommended Product" probably shouldn't have stochasticity. This problem opens up some interesting ML cans of worms, but I believe these issues could be overcome.
3) Another issue was if I wanted to scroll in the left column while the right column was still loading, the scrolling was very janky. This was an issue on firefox, but it took quite a long time for the app to be functional (> 10s)
4) Perhaps you could move the search bar and the logo to the top, so the logo is on the top left corner and the search bar takes space to the right of it. This way there aren't overlapping elements, I'm sure there's some annoying edge cases there which would frustrate users
5) For negative ingredients (and maybe any of the ingredients) it would be nice if you kept track of an ingredient database with references. I want to know why some ingredient is bad for my skin, and what I could expect.
6) If a product has many distributors, my first through was the arrow scrolling through products was a slider for the distributor list. I wonder if there's a nice way to differentiate the arrow further, so its functionality is more apparent.
Anyway, this is an excellent proof of concept, I'm excited to see how this product develops.
As for the performance issues, we're looking into several things that could speed things up
- Fine-tuning a small LLM for the results on the intermediate page and deploying on a provider with higher throughput and time to first token
- Admittedly, there there are quite a few SQL query / index optimizations we need to make on the backend, along with making parts of our pipeline async
- The frontend itself is also not very performant right now, but we're working on it.
We cache previous calls to the API, so that's why the demo queries or queries others have tried before you are faster. I'll ship a change that makes the results more consistent but not fully consistent later today.
As for the ingredients, citing sources is definitely a next step. In the meantime, I recommend looking up the ingredients that catch your eye on a place like EWG Skin Deep if it's a huge concern for you (I used to do this to make sure my ingredients weren't comedogenic for acne).
Great point about the distributor list UI, we'll think about a better way to show it!
Congrats on the launch! I'm excited to see how your product develops.
A quick note: while playing around with the postmortem report presentation, I ran into a white screen of death. I appreciate you gave the user an error message, many folks just let the white screen appear and that's it! Anyway, if you'd like to reproduce the crash, here's what you do:
1. Open up the presentation mode by clicking on the "> Present" button in the top right corner
2. Navigate to the first graph page
3. Click the settings icon
4. Change the graph type to a bar graph
5. Under the "X" coordinate section of the graph settings, change the dropdown from "TIMESTAMP" to "100*(...)"
6. This will give an error message stating "Application error: a client-side exception has occurred (see the browser console for more information)."
and in the console I'm given a message saying
TypeError: r is null 69-b1345c3baf602dab.js:1:3791
maybe that's useful?
Anyway, congratulations again, I'm excited to see Thorntale's future progress!
Hey, this is a neat looking tool! I have a few thoughts about QA issues I ran into while playing around with Krata Maps:
1) When I went to create a polygon and put down a few points, I wasn't sure how to finish dropping points. I tried a few different strategies, and ended up accidentally deleting the polygon, before I found out the correct way to do it. First attempt was clicking on the black part of the top toolbar, but this did nothing. Second attempt was clicking on the mouse pointer tool, but that cleared the input. Third attempt was right clicking, but that cleared the polygon tool. Now I'm not someone who typically uses keyboard shortcuts when first using a product, so pressing "Enter" was the last thought on my mind. Before I tried "Enter" I made the mistake of pressing "Esc": shoot, another cancelled polygon.
Note that when using Figma, they provide a "Done" button while using the SVG tool. This was a bit of a frustrating experience to not only be stuck without a polygon on the page, but the app ended up deleting my work upto that point.
If someone was using this tool to draw something intricate and made a slip of their hand near the end, oof, that would make me want to rage-quit computers for the day.
2) Also, there was not undo functionality. It would be nice if I made a mistake and could press ctrl-z or press an undo button to get back into using the polygon tool with the previously placed dots put back in their correct spot
3) I think it's confusing a polygon isn't added to the features column as soon as you put down the first node.
This has some strange side-effects: one is I can click on other features, edit their properties, but still be in the polygon editing mode for the polygon which hasn't been added to the features yet. Another is I can click around and edit features, but when my mouse goes back to the map, it's still in polygon editing mode.
Again, I love this product concept and will keep a look out for updates. It has a slick looking UI and is a really cool concept. Good luck!
This is wonderful feedback! Will try to fix some of these soon, thanks. The project was built around 2 years ago, but it failed in giving any signs of generating revenue. It was abandoned for a while, I shut down the cloud version and open sourced part of the codebase.
Great job launching your product! This interface looks slick and I'm excited to see your product progress. Hopefully I can get beta access ;)
I have some QA feedback for some quirks I noticed and some potential feature ideas/requests:
1) On the companies list, it would be nice if you automatically grabbed the favicon.ico from the domain. It would be nice if you could tether fields together so information can be automatically derived from other fields. This could include an area for custom scripts to do this automatically.
2) While using the demo I was automatically logged out. I'm wondering if someone else logged in and then my session was killed? Also, when I was logged out, I was given a white screen of death when I clicked back on the history, but was given no redirect. E.g. pressing back sent me to https://app.okapicrm.com/objects/companies but I'm just stuck with a page with the title "Title" and an empty div[id="react-root"] node.
3) Because you end the password with a period, if I double click on the password and copy that, it will put the wrong password in my clipboard. I wasn't sure if this period was part of the password or not until I logged in and tried without the period
4) At first, it wasn't obvious the search feature was only available for the current page you were on. E.g. I tried accessing search on the reminders page, https://app.okapicrm.com/reminders, and I wasn't sure what was suppose to happen after I typed something in. There was no visual feedback that a search query was being processed, or had been processed. Once I tried it on another page was it clear that the search feature was searching for records visible on the current page. Also, because it is so quick, I wonder if this is purely a client-side search. If I have a Records page with a paginated number of records, will the search be able to find records which are not visible on the current page? If not, I would find this super frustrating as a user.
6) I'm not sure I understand why fields which point to an archived record must be cleared. Could you keep those fields and just use some kind of visual indicator (such as a red highlight or background) to show the record being pointed at is archived?
7) When I clicked the archive button, there was no visual feedback the website was loading. It would have been nice to at least change the cursor over the modal to a "cursor: wait" while it was making the archive request.
1. Definitely want to improve favicons. For deriving fields, we do have formula fields. Would love to show that to you, it's not obvious enough how to use to enable for this HN launch.
2/3. Sorry about that. This username/password approach for the HN launch is definitely a hack.
4. Right now, search is just on the current type of record you're looking at. I want to fix this, and I just haven't because I haven't made the UI for filtering what objects to search across. The backend is in place though. Glad it seems suspiciously fast! I can assure it's a real FTS talking to a database.
5. I want to fix this too. A way to make data of type "Link" for better UI treatment.
6. Archiving is meant for data that others shouldn't be looking at -- it's a staging ground for data. That's why you can't point into archived data. It's also a precursor for deleting data, and we want to guarantee no dangling references to deleted data.
Feel free to shoot me an email (found in my profile) when you update the product, I'm curious about how the formula fields will work and what kind of interface I'll be given.
Could you give more detail as to how it generates the book? For example, does the user write an outline? Or is it just writing a book based on the title alone? It's not very clear to me what spending $99 is giving me.
Ok so user enters book title like "Ultimate Guide To Europe Travel", then select number of chapters (say 6), select langauge (English, German, Italian, Spanish) and then hit generate book. It takes around 9 ~ 15 mins. The AI generates the book outline based on the book title and number of chapters. We are shipping the custom Chatper, Subchapter functionality later this month/early next (requested by users).
QAComet is a one-man QA As A Service agency providing unlimited QA requests for $5k/month. There is a $50 14 day trial for the first three customers who contact me through HN. My services include
- E2E/Integration/Unit testing
- QA/Usability reviews
- Bug finding/reporting
You can see a case study of some past work I've done with Automatisch.io helping build their E2E test suite and other QA work here
- https://qacomet.com/case-studies/automatisch
Also, I am currently working with MedusaJS helping get their template project ready with a suite of automated tests. Some example bug reports include
- https://github.com/medusajs/nextjs-starter-medusa/issues/295
- https://github.com/medusajs/nextjs-starter-medusa/issues/269
Here's some more features my service includes
- $50 14 day trial
- 1 request at a time
- Unlimited requests 24/7
- Async workflow
- Pause subscription at any time
Email: lucas@qacomet.com
Website: https://qacomet.com