Never found a good application for any of these immutable js data libraries. Maybe it could be helpful to prevent bugs in middleware chains of expresslike backends but for ui libraries it always felt like the libraries basically recreate mutable data structures with virtual proxy objects but at that point what does immutable even mean? you have completely virtual objects and can do whatever you want to make the rendering performant and reliable, at that point it would look probably more like a data binding layer with virtual objects to something like tanstack query
I still don't understand how people consider bun as a viable runtime when its owned by evil corp trying to use it for capture of the tooling layer and its the most insecure runtime on top of that. Meanwhile deno is performant and has dramatically improved node compatibility while exposing a proper permission broker api.
Same reason linux or databases need to be open source. A sandbox is not a nice to have or a feature anymore, it is as fundamental building block to running any software. You cannot depend on closed source building blocks, not as closed source product and especially not as open source product.
Not really, its more like saying no one needs another windows when linux exists. By "no one needs" i mean the world needs open source sandbox building blocks that are up to the challenges of the current age, no closed source solution can be a fundamental building block for the world to become better and more secure. No non-local building block can be at the foundation to anything that makes the world better and more robust.
I had to dig hard to find this is a SAAS sandbox offering not an actual sandbox (the software i can use locally). Its just wasting peoples time, no one needs a non opensource sandbox. There are now at least 3 apache 2 projects (smolmachines, microsandbox, boxlite) working on sandboxes and at least one of them should be ready for primetime soon.
We still have not the right sandbox and PR abstractions to make the merge of the two complete. Imagine merging a PR and knowing exactly this code cannot ever possibly reach the internet and it can only receive and send specific shapes of api requests from these specific services, it has well defined resource limits and you have specific optimal UI to review these constraints. I can imagine to not review a bigger number of PRs in that reality.
Most software has a few years lifetime and nearly no users. What you say is only true after reaching a certain milestone like product market fit. I think the idea is to reach that turning point as fast as possible and then rebuild the system from ground up with maintainability and quality focus.
Bun is basically a lost case in the same way that claude code is. Many people love both and thats fine as long it is their personal choice. But freedom loving orgs can never ever depend on either for anything ever or build anything on top. The main problem in this picture is projects who don’t understand or remember the internet explorer hell are using bun only features to a degree that you cannot just switch. Even opencode who should have the biggest aversion to depend on anthropic is doing this and it drives me insane. Have you all lost your mind or let your agents blindly lock you into whatever hell wants to own you?
Its completely impossible to trust an article when half the diagrams have serious issues and headlines are worst slop indicators. I am sure the author is knowledgeable and spend some time on this. But please please, with butter, please, either spend the last 10% to get rid of these issues or just don't publish it! Keep it in your ai research gallery or if you want to publish that gallery as ai research?
An aspect i don't hear about enough is using agents to make better abstractions instead of treating the agent as the abstraction. We could work on making abstractions that allow expressing what we are building in most concise form without any boilerplate. The reviews are enjoyable, if llm providers fail we have a chance to use the great abstractions by hand. The alternative i see everywhere: We don’t care about the bad abstractions because agents write the code. We don't review because its too much and too noisy and we rather blow our brains out than manually dig through that react and spaghetti codebase and writing all the boilerplate and cross references by hand. We have to build for resilience and future us.
The title is highly misleading, they mean the harness belongs outside the sandbox the agent is working in. Please run also the harness in a sandbox, i don't think any of them is safe for a host. This is also the only valuable info in this marketing noise article. The rest is full of hidden endorsements of VC buddies (why would anyone build on closed source sandbox abstractions and claim alternatives need 1s to boot) And signs they cannot reason logically. (eg. Need shared files across users > the only two options are building distr. filesystem or storing in a database. Later admitting even the database solution needs a last write wins resolution on top and completely ignoring they could just as well have delegated shared writes to a authoritative file server with retry on conflicts and git.)
I don't think this should be dismissed as a cheap and rude ripoff. I'm no expert in trademarks or the naming convention part of the story, but for the rest: warp is not a great company taking far too long to roll back its weird account requirement, tracking users, enshittifying the core terminal experience with often unwanted AI and other crap features and dismissing annoyances that are brought up by 100eds of users. We need to show companies they need to behave or will be crushed by the community.
apart from not giving the "neither" option of turquoise, you cannot just line up asking without neutralising perception in between eg. with a non blue/green and then white before the next test. Color perception is >relative< not just across individuals, not sure how someone interested enough in blue perception does not understand this.
This post is a gem. I wanted to build something like this for 10 years but shied away until last year because your journey seemed exactly what would happen going down that road in a world not ready yet.
I started with regular valve inspired "playtesting" with users on prototypes which showed:
- Power users love canvas and multiplayer but only need it in distinct product or company phases which last a few weeks or months max, outside these phases they need that maybe once every one or two weeks which is not enough to stick to a new browser
- outside of the usage, they feel burdened by mental overhead of collaboration or canvas features and much prefer simplicity
- the world does not need another closed source browser, VC is fundamentally incompatible with what is needed, so any step after joining yc was doomed from the start
Im really tired of people saying "the agent did this" or posting agents excuses as if they still think agents behaviour is a safety layer not a mere usability tool. Posts like this reinforce this misunderstanding in juniors instead of learning to focus on the workflows and tools. "well, you should have used a better model." >> this is nothing any sane person even remotely knowledgable will ever say. Non deterministic systems gonna nondeterminist so what? The issue is relying on cli/imperative tools and seeing manual changes to subdomains as a casual, when in reality there are a lot of implications on changing your domains (or anything about your hosting setup), this should be completely automatic and the system to do this needs be driven by gitops with declarative truth, you know the things the devops world has been perfecting and saying for the past 10 years?
The only missing interesting thing is: did this token file live inside the current project folder? Or did cursor fully fail to constrain actions to the sane default? In either case i make a strong point to disallow agents accessing any git ignored files even if inside the folder, this will prevent a whole breadth of similar problems, with minimal downside, plus you can always opt subsets of ignores back in where it makes sense.
One last point i want to make is do not trust just your agent harness, if it matters at least require one or more layers of safety around the harness. Use sandboxes or runtime enforcement of rules. Do not accumulate state there but use fresh environments for every session. This will reduce the risk for things like this happening by an order of magnitude.
im just about giving OPs premise another go. compose just feels so much better as abstraction especially with small and medium setups looking close to the optimum of expressiveness without boilerplate to describe what is needed. The missing pieces seem to also be in the compose compatible “docker stack” aka new docker swarm, which i ignored for probably too long as i assumed it was the discontinued old swarm. Even if new swarm mode sucks how hard can it be to make something compose shaped vs running k8?
All these agent memory systems seem so simultaneously over and under engineered and like a certain dead end. I cannot imagine any reality in which this does not rot and get out of sync with what the latest model need. For the one time you build a payment provider how many session will be tilted towards thinking about payments because of the "don't use stripe" memory?
I have trouble seeing how this is different to linode, if i invest time in a new VM api, this has to work for cloud or my own machines transparently. Lastly as much as i share the disappointment in k8s promise, this seems a bit too simple, there is a reason homelabs mostly standardised on compose files.
You have to look at it from the workflow side not compare the tool on your current git workflow. Check out https://www.stacking.dev for a description of stacked commits, which is more the everyday workflow on jj than the octopus in this article.
I don't see storing non-sensitive environment variables unencrypted as the main issue here. Sure at vercels scale, encryption at rest for any data would add some better baseline, but i see this article as two major user interface fails more than anything else. Oauth dialogs are just pathetic, they are years behind what is required and what UX research knows how to do things, none of the companies invested any amount of resources into it after it just worked well enough not to make most users churn. The env var problem is also ridiculous, you can only update, not see and check values in the interface if they are encrypted for most providers i know, that leads to really annoying UX and is the reason they are not marked as sensitive by default and opt out. Even if you could unlock them to edit, no one will enter their password again as that is too much hassle, meaning we need a way to read and edit encrypted env vars in the interface where they are created but not have more in the way than a passkey dialog. Its doable but afaik no provider would go the extra mile to get to this UX.
(Of course there are tons of other red flags not looked at in the article, eg. how does an employees machine get access to production systems and from there access to customers connected with oauth and how does the attacker get to env vars from a google workspace account)
I only know pausing games is funky because the highest my playstation fans ever go is pausing some games. Quite weird pausing is not just a feature of the game engine or runtime, especially as the menu and settings system seem to be totally separate in most cases anyways.