What's with the pricing of these sandbox offerings recently? I assume just trying to milk the AI trend.
It's about 10x what a normal VM would cost at a more affordable hoster. So you better have it run only 10% of the time or you're just paying more for something more constrained.
A full month of runtime would be about $50 bucks for a 2vCPU 1GB RAM 10GB SSD mini-VM that you can get easily for $5 elsewhere.
We have also a service impacted by this. There is no status acknowledgement of the severe degradation. Most of the time queries just take 2-10s, some even 20s+ and a few times they even hang indefinitely. It's a CF-internal networking issue and not SQL execution issue because in the meta data you can see SQL processing taking like 1ms. Even a simple "SELECT 123" in their dashboard D1 console can take 10+ seconds.
Our company opened an issue but we have not heard back in 3 days. The issue has been going on for more than the 3 days, we've noticed it nearly a week ago.
You can also find people in the CF community forums or in other places complaining about D1 latency and networking issues going back as much as a year. Nobody from CF replies in their community forums.
I think this is 1. not an acceptable level of service 2. not acceptable level of support 3. very surprising for a company who is all about fast and reliable networking.
Unfortunately I experienced all kinds of problems with Cloudflare and their non-CDN services over the past couple of months and I've also never seen a major cloud service provider suffering from so many dashboard errors and bugs.
Cloudflare folks, if you read this: you need a Snow Leopard year. You've launched many services and products but they are riddled with bugs, performance issues and outright outages. The recent launch of your data warehouse offering which in our testing (and I think even in your presentation) showed data processing speeds measured in the kilobytes/s while not even supporting aggregations at launch would have been embarassing even 20 years ago.
Please please step up your quality game because you do have some interesting offerings with good potential. The two huge outages last year were just surfacing deeper issues that seem to permeate throughout the stack.
Cloudflare Containers (and therefore Sandbox) pricing is way too expensive. The pricing is a bit cumbersome to understand by being inconsistent with pricing of other Cloudflare products in terms of units and split between memory, cpu and disk instead of combined per instance. The worst is that it is given in these tiny fractions per second.
Memory: $0.0000025 per additional GiB-second
vCPU: $0.000020 per additional vCPU-second
Disk: $0.00000007 per additional GB-second
The smaller instance types have super low processing power by getting a fraction of a vCPU. But if you calculate the monthly cost then it comes to:
Memory: $6.48 per GB
vCPU: $51.84 per vCPU (!!!)
Disk: $0.18 per GB
These prices are more expensive than the already expensive prices of the big cloud providers. For example a t2d-standard-2 on GCP with 2 vCPUs and 8GB with 16GB storage would cost $63.28 per month while the standard-3 instance on CF would cost a whopping $51.84 + $103.68 + $2.90 = $158.42, about 2.5x the price.
Cloudflare Containers also don't have peristent storage and are by design intended to shut down if not used but I could then also go for a spot vm on GCP which would bring the price down to $9.27 which is less than 6% of the CF container cost and I get persistent storage plus a ton of other features on top.
You can’t compare these with regular VM of aws or gcp. VM are expected to boot up in milliseconds and can be stopped/killed in milliseconds. You are charged per second of usage. The sandboxes are ephemeral and meant for AI coding agents. Typical sandboxes run less than 30 mins session. The premium is for the flexibility it comes with.
I think you can absolutely compare them and there is no added flexibility, in fact there is less flexibility. There is added convenience though.
For the huge factor in price difference you can keep spare spot VMs on GCP idle and warm all the time and still be an order of magnitude cheaper. You have more features and flexibility with these. You can also discard them at will, they are not charged per month. Pricing granularity in GCP is per second (with 1min minimum) and you can fire up firecracker VMs within milliseconds as another commenter pointed out.
Cloudflare Sandbox have less functionality at a significantly increased price. The tradeoff is simplicity because they are more focused for a specific use case for which they don't need additional configuration or tooling. The downside is that they can't do everything a proper VM can do.
It's a fair tradeoff but I argue the price difference is very much out of balance. But then again it seems to be a feature primarily going after AI companies and there is infinite VC money to burn at the moment.
I coud easily spin-up a firecracker VM on-demand and put it behind an API. It boots up in under 200 milliseconds. and I get to control it however I wish to. And also, all costs are under my control.
In my case, it is ignorance. I am not familiar with how to wield firecracker VMs and manage their lifecycle without putting a hole in my pocket. These sandbox services(e2b, Daytona, Vercel, etc.) package them in an intuitive SDK for me to consume in my application. Since the sandboxing is not the main differentiator for me, I am okay to leverage the external providers to fill in for me.
That said, I will be grateful if you can point me to right resources on how to do this myself :)
This is a pretty good use-case for an open-source project then.
For guide, just follow their official docs. I did those again today, literally copy-pasted shell commands one after the other, and voila.. had firecracker vm running and booting a full-fledge ubuntu vm.
It was sooo damn fast that when it started, at that moment I thought that my terminal had crashed because it's prompt changed. But nop. It was just that fast that even while literally looking at it I was not able to catch when it actually did boot-up.
By the way, two open-source projects already exist:
Cloudflare containers feel a lot more pricey as compared to workers but I think that it could provide more streamlined experience imo but still, If we are talking about complete cost analysis, sometimes I wonder how much cf containers vs workers vs hetzner/dedicated/shared vps / gcp etc. would work out for the same thing.
Honestly, the more I think about it, my ease of sanity either wants me to use hetzner/others for golang/other binary related stuff and for the frontend to use cf workers with sveltekit
That way we could have the best in both worlds and probably glue together somethings using proto-buf or something but I guess people don't like managing two codebases but I think that sveltekit is a pleasure to work with and can easily be learnt by anybody in 3-4 weeks and maybe some more for golang but yeah I might look more into cf containers/gcp or whatever but my heart wants hetzner for backend with golang if need be and to try to extract as much juice as I can in cf workers with sveltekit in the meanwhile.
Isn't that a bit of a holy grail though? If your software can fact check the output of LLMs and prevent hallucinations then why not use that as the AI to get the answers in the first place?
Because you - hopefully - have a check against something that is on average of higher quality than the combined input of an LLM.
I'm not sure if this can work or not but it would be nice to see a trial, you could probably do this by hand if you wanted to by breaking up the answer from an LLM into factoids and then to check each of those individually, and to assign a score to them based on the amount of supporting evidence for the factoid. I'd love that as a plug-in to a browser too.
I don't think they are a hint that spacetime is not fundamental. But I do think spacetime has to be some kind of real physical reality.
The modifications of spacetime that we see as effects of gravity are relative changes to our immediate surroundings or reference frame.
Similarly how you can't tell who is actually stationary and who is moving when two objects are in freefall and all you can note is the relative speed between the two, it would be equally valid to say the objects inside spacetime are getting distorted relative to spacetime.
Even on my Macbook with Firefox the site has a strange feel when scrolling. It's not exactly struggling but it feels unnatural and slightly off/slow/uneven. Like it's on the edge of struggling. Bit hard to describe. The effect gets worse towards the mid section of the page with the side scrolling logo circles. I removed that section via dev tools which helped with performance. When I have that part of the page in view I get 80-90% CPU usage of one core. But even after removing it I can saturate a core by scrolling around, especially towards the lower part of the page.
It is indeed one of the worst optimized CSS I've seen in a while. Weird for a project that is all about speed.
If every site did that then it would be harder to quickly spot one in a long list of tabs. A neat trick but I don't think it is a particularily good idea.
It doesn’t say the paper was the entity engaged in the raid. If I didn’t know the broader context, I would assume that sentence meant “A newspaper was investigating a police chief at the time the police chief’s home was raided by another law enforcement agency”. That seems way more likely than a newspaper being referred to as “him”.
There is some meat to the story, I agree. But it's not surprising. The fine tuning model of course will be small in file size and not take too long to train because by definition it is applying changes to a small subset of the main model and is trained only on a small amount if input data. You can't use the small tuning model for "Teddies" with a query that has nothing to do with Teddies. You could see these small tuning models as a diff file for the main model. And depending on the user query one can choose an appropriate diff to be applied to improve the result for that specific query.
When you train a model with new inputs to fine tune you can save the weights that got changed to a separate file instead of the main file.
In other words one can see the small tuning models as selectively to be applied updates/patches.
It's about 10x what a normal VM would cost at a more affordable hoster. So you better have it run only 10% of the time or you're just paying more for something more constrained.
A full month of runtime would be about $50 bucks for a 2vCPU 1GB RAM 10GB SSD mini-VM that you can get easily for $5 elsewhere.
reply