peter_l_downs

Thank you!

Thank you

Thank you — from that page, at the bottom, I was able to find this link to what I think are the quantized versions

https://huggingface.co/NousResearch/Hermes-2-Pro-Mistral-7B-...

If you have the time, could you explain what you mean by "Q5 is minimum"? Did you determine that by trying the different models and finding this one is best, or did someone else do that evaluation, or is that just generally accepted knowledge? Sorry, I find this whole ecosystem quite confusing still, but I'm very new and that's not your problem.

Sorry to repeat the same question I just asked the other commenter in this thread, but could you link the model page and recommend a specific level of quantization for the models you've referenced? I'd love to play with these models and see what you're talking about.

I'm just starting to get into downloading and testing models using llama.cpp and I'm curious which model you're actually using, since they seem to come in varying levels of quantization. Is this [0] the model page for the one you're using, or should I be looking somewhere else? What is the actual file name of the model you're using?

[0] https://huggingface.co/TheBloke/OpenHermes-2.5-Mistral-7B-GG...

Selfishly, I'd love a documentation review if you (or anyone else) has the time to take a try out some golang projects I've been working on.

The code is frankly not very polished and not worth reviewing, but I'm curious to hear feedback on the documentation / README and whether or not you clearly understand what these libraries are for and how to use them.

* https://github.com/peterldowns/pgtestdb

* https://github.com/peterldowns/testy

* https://github.com/peterldowns/pgmigrate

Ah, that makes sense. Thanks!

This isn't supposed to happen, the repo is a read-only mirror. Very weird. Anyone know what's going on?

edit: maybe an april fools thing?

> On top of that, Robinhood is on the hook for 5% APY on uninvested cash. Yes, the uninvested balance is cash that Robinhood can invest and profit from, but will they clear the 5% hurdle every time?

Just one example, of many, that shows this author doesn’t know what they’re talking about. At all.

For free, you could:

- Borrow a physical or digital copy from your library

- Borrow a digital copy from the Internet Archive: https://archive.org/details/fasttimesatridge0000crow

- Download a digital copy from Anna's Archive: https://annas-archive.org/search?index=&q=fast+times+at+ridg...

Enjoy!

Ahhhhhh of course. Thanks, that makes sense.

Yes, to every single one of your questions.

That's astonishing. What ended up happening?

Quotidian viewpoint that everyone working in search already agrees with, for years, but nice to see it written down. Embedding-only search has never survived its first contact with an actual user.

And the documentation makes me think they did a great job making this easy to use. Looking forward to playing around with it.

Edit: oh you’re one of the authors — thank you, and congratulations!

I had an accessibility issue with Github today and I attempted to file a bug report. It was a lot harder to do than expected, and confusing, and riddled with spam. Has it always been like this?

Here's what I went through:

1. Notice the problem (tooltips showing up and making the click targets for repository names confusingly and annoyingly small.)

2. Click in the "upper left hamburger" menu to try to find a bug report page. Didn't see anything relevant.

3. Click in the "upper right user icon menu" to try to find a bug report page. See two relevant links, "Github Support" and "Github Community".

4. I try "Github Support", and I have to log in "with Github", to access a subdomain support.github.com. Weird.

5. Search for "tooltip" to see if this is a known issue. See some results, click through them, end up on the "github.com/orgs/commmunity" site where people ask questions and seem to report bugs.

6. Report the bug, here https://github.com/orgs/community/discussions/114049

7. Click on the main "discussions" tab to see other reported bugs. Home page is literally filled with spam.

8. Try to search for tickets reporting the problem of spam. Find this existing ticket, and add my own report of the spam problem in response to the "sorry there's an active phishing campaign going on right now" answer https://github.com/orgs/community/discussions/111562

9. Get a bad feeling in my stomach. There's no way this is the expected way to file a bug report, I pay for this software. Go back to step 3 and try "Github Support" again.

10. Search for "how to file a bug report" and find this answer https://github.com/orgs/community/discussions/23084 which links me to https://support.github.com/request and the "bug report" option.

11. Finally file an actual bug report.

I'm beyond confused. How is this so bad? Every step in the process, just, terrible. I checked back in and noticed even more spam. Figured I'd report it here in case that helps get it looked at faster. What a terrible product.

I don't understand what you mean, what is "node extra ca cert etc" and what is the issue with mkcert?

Localias wraps Caddy to handle all the cert provision; I believe Caddy uses mkcert. I haven't seen any bug reports about yet, but if you give it a try and run into an issue I would be happy to help fix it.

I'm glad it worked for you and was easy to set up. If you run into any trouble or have any feature requests, please file an issue on github!

Yes! Thank you for making Caddy, it’s wonderful software and it was really easy to extend!

I like the interactive setup. I think this is solid but if you want something even faster and easier to use, try my project localias [0]. The parent project, lcl.host, has some annoying restrictions:

> This CA has some restrictions though: it can only issue certificates for subdomains of lcl.host and localhost, but that’s all you need for local development.

Localias, on the other hand, lets you use any custom domain you'd like. And if you use a domain ending in .local, it will broadcast over mDNS so that you can easily connect to that server from any other device on your wifi network (like your phone.)

Localias also allows you to share your configuration with your entire development team by committing a .localias.yaml file to the root of your git repo. This makes sharing links with each other super convenient.

Always nice to see another competitor in the space; if you're interested in this, please check out Localias as well!

[0] https://github.com/peterldowns/localias

I view a comment like yours as essentially spam and of negative value. I'm not sure if we (vaguely, hn forum community) have agreed on the etiquette here, but I have to ask why you would bother sharing an AI-generated response if you're not sure if it's correct. What are you getting at? How could this be helpful?

I was having trouble finding the link to the actual project, so here you go:

- Github repo https://github.com/Caranea/gpulife

- Finished particle simulator https://webgpulife.netlify.app/

You’re right, the point I was trying to make but should have said more explicitly is that they raised at an inflated valuation that would make any subsequent “up” round more difficult to raise in the future, even if the growth trajectory improved.

Yes, they care, although this of course varies. Remember that in this situation: one of the founders had already left, another was "giving up", and other key employees had left. Tech companies were laying off like crazy, and your chance of getting your money back via another round was about 0%. Would you be willing to subsidize a bunch of jobs in a bad-trajectory company with disinterested/absent leadership? Or would you instead think to engineer an outcome like this where the employees get a job offer and a bonus, one of your other investments gets a chance at top talent, a serial entrepreneur founder remains happy with you, and you get some money back?

Depending on timing (and here it was just over a year between the round closing and the acquisition), you may be able to easily re-deploy the capital you get back as another investment included in the same fund, and get another shot at a 100x return.

I'd be curious to hear what others think (particularly actual VCs, not just pretend ones like me) but this seems like a very rational outcome to me.

> I imagine they would have much more preferred us to continue working hard than to return the bit of capital.

You imagine incorrectly, as revealed by your lead investors' decision to allow an acquihire by another one of their portfolio companies.

Even if Airplane had been cash flow positive, slowing growth and a demotivated CEO and departure of key employees at an early stage is a tough pill for investors to swallow. See my other comment, but I find it highly likely that your investors were pushing for a way to wind things down and return as much of their cash as they could get, and an acquihire is a fairly employee-friendly way to do that.

I would recommend either updating your framework to allow for instrumentation (logs/metrics/etc) or showing how to add that instrumentation in "user defined workflow states" via an example application.

The current pitch, which is "this is a workflow orchestration framework that does not allow for any monitoring or observability" is a complete non-starter. You may want to consider how other projects keep track of workflow state and allow for it to be instrumented.

Thank you for clarifying, no problem — always fun to learn something new.

Is this the NATS you're referring to here? Forgive the noob question, I just haven't come across that acronym before (except in the Network Address Translation context.) https://nats.io/

Looks like a great side project, congrats on launching. My basic feedback is that you should set expectations on the project page by answering the following questions:

- Why would someone use this instead of Airflow/Cadence/Temporal/Databuilderframework?

- What does this look like when it's used? Most frameworks provide some kind of example project, you should too.

- Related, but more specifically, what does the `IDataStore` interface contract mean? Beyond the two functions that I have to implement, are there any considerations related to the overall performance/scalability/durability of the system? Would it make sense to use a disk-backed store, or Redis, or Postgres?

- How do I observe the system? Which workflows are running, which have failed, what the current state is, etc. Are there metrics? Logs?

All of this is based on the assumption you want people to adopt this framework. If it's just a cool side project, that's fine too, but you should probably try to set that expectation in the README.