Hacker News new | past | comments | ask | show | jobs | submit | rita3ko's comments login

(confession: i work at cloudflare)

but we just made our first announcements for this week, including: - d1 (our serverless database) going GA - hyperdrive (SQL query cache for postgres and mysql) going GA - and workers analytics engine

if you just moved off google domains, i recommend checking out cloudflare pages, which is cloudflare's product for hosting static and dynamic sites (with gitops, everything else included).


Yup, I'm planning to check out Pages for some static sites. R2 might be another one I look at.

I'm largely container / k8s based, no real interest in these edge or serverless solutions that cannot be easily replicated locally. I'm also keen to keep intra app requests in the same cloud provider, to avoid latency and egress costs

("disclaimer" is probably better than "confession" Cloudflare is generally held down on high regard here :)


you can do both! all of our models are supported either via workers / pages binding (which makes it really easy to host the rest of the logic), and via REST API

docs: https://developers.cloudflare.com/workers-ai/get-started/res...

(llama specific example here too under curl: https://developers.cloudflare.com/workers-ai/models/llm/ )


Awesome! Thanks for that. Cloudflare smashing it on simplicity as per usual.


we always are! email is rita at cloudflare dot com :)


thanks!


stay tuned!


Any chance you're looking for technical product folks to work on this? I actually lead a very similar deployment internally at Livepeer (focus was on live video enhancements but also generalized edge compute)!


You can always email jgc@cloudflare.com and I'll route the resume to the right people.


thanks!


we (cloudflare) have a great example of a vector database build on workers + kv here: https://ai.cloudflare.com/demos/vector-store/

(plan to talk about how we built it soon)

edit: clarify i work at cloudflare


> plan to talk about how we built it soon

Neat! Reminds me of using DynamoDB as a multi-dimensional store with (space-filling curve) z-order indexing [0], which isn't as fast (as native spatial databases, say) but it works. I wonder how performant retrofitting Workers KV is compared to a native vector database? Please do publish benchmarks, if possible.

[0] https://archive.is/ZGY1T / https://news.ycombinator.com/item?id=30009893


Is this open source somewhere? It seems interesting to study


What details are you looking for? Maybe we can fill you in :)

We still have some work to do to get the runtime ready to open source and make it broadly available, but happy to answer any questions we can.



Pages itself has been GA for a little while now (https://blog.cloudflare.com/cloudflare-pages-ga/). For the integrations with Workers, it's TBD (we generally want to leave things in beta for a bit and let it bake in before giving you the thumbs up to put your production site on it), but we move fast, so likely sometime in the first half of next year.


hi there! i'm on the workers team at cloudflare.

you're right that the ideal application would run fully on the edge (which is why we try to offer things like KV, Durable Objects, and more options in the future for you to have your data where your logic is running).

however — moving your data to the edge is unlikely to happen overnight. and it's going to take us a bit to offer you all the data options for all the different use cases that you need (after all, there are many types of databases for a reason), so this gives users greater optionality. secondly, if your data is already somewhere but you want to build new features on top of it, now you can. (say, i already have a database with all my apartment listings in it, but i want to add a new feature where i let my users add photos of the apts). i can now build that feature in workers, without having to undergo an otherwise massive migration.

one last thing, this is just the beginning of what even database connectors mean. in the future, we can cache data, and actually improve performance and reliability even if the dc is far (as you said, make it easier to have best of both worlds).


Thanks for the response! I hadn't considered that last trick. If you can make it work well, it will be pretty cool.


also a good time to mention that if you build something in in rust using workers, you should submit it to win some free swag:

https://challenge.developers.cloudflare.com/


Consider applying for YC's W25 batch! Applications are open till Nov 12.

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

Search: