You phone and computer will be checked thoroughly using automated tools. If they didn't find any sensitive keywords, you'll be fined and recorded in the system. If they find something, a detention for at least 3 days or ... forever.
> It's fascinating that browsers are one of the most robust and widely available sandboxing system and we are yet to make a claude-code/gemini-cli like agent that runs inside the browser.
It's easily explained by the fact that all the javascript code is exposed in a browser and all the network connections are trivially inspectable and blockable. It's much harder to collect data and do shady things with that level of inspectability. And it's much harder to ban alternative clients for the main paid offer. Especially if AI companies want to leave the door open to pushing ads to your conversations.
My wife and I are experimenting with a small side project: T-shirts inspired by vintage covers of classical novels we love. The idea started as "would we actually wear this?" and turned into a tiny collection.
We even made 1 sale, the customer loved it, that made us very happy and motivated :)
We're still early and very much in exploration mode, so I'd love feedback. Anything I should do add/remove to the website also counts. :)
If you have favorite classic novels, cover designs, or sites/resources about vintage book art that you think we should look at, I'm all ears.
Something that I always wondered about this tool, and why I bucketed it in my mind as a "cool but ultimately boring toy" - why are all effects perfectly spherical? Surely the actual effect, especially for lower-yield devices, would be meaningfully affected by local topography and even buildings?
Yes, this is basically what happens. Except sometimes there's no realistic way to narrow the estimate. The idea that you should "scrap" a project that you can't break down doesn't carry over to research-heavy teams. Such teams need to have a way to manage wide estimate windows.
Accusing people you disagree with of being "very far right" to automatically discredit them withotu arguemnts, is the ultimate bad faith cheat code of online arguments. If you want, we can have this conversation over another medium where I can share you the data from government sources that prove my point as being mathematically and logically sound, and not "far right". THere's no point continuing here since HN bans such discussions as inflammatory without right to appeal, regardless of what data/arguments you bring to the table so even if you win, you still loose.
> I can tell you based on your description that you did not do this. Subagents are completely different and cannot be used in this way.
And yet I have used them exactly in the way I described. That you assume they can't just demonstrate that you haven't tried very hard.
> No, it isn't how Claude Code works because Claude Code is designed to work with limited task queues, this is not what this feature is.
Claude allows your setup to execute arbitrary code that gets injected into context. The entire point is that you don't need to rely on built in capabilities of Claude Code to do any of this.
> No, it doesn't work within that context. Again: sharing context between subagents, single instance running for months...I am not even sure why someone would think this could work.
I know what I described works because I am doing it. You can achieve what I described in a variety of ways: Using skills to tell the agents how to access a shared communications channel. Using MCP servers. Just using CLAUDE.md and describe how to use files as a shared communications channel.
This is only difficult if you lack imagination.
> You are talking about having some CLAUDE.md files like you have invented the wheel, lol. HN is great.
No, the exact opposite: I'm saying that this isn't hard, that it isn't anything revolutionary or even special. It's pretty basic usage of the existing facilities. There's no invention there.
You're the one trying to imply this is more revolutionary than it is.
The bindings are still done at the JS level. But to answer your question, I'm building a git workflow engine (kind of a lightweight GitHub Actions alternative; see https://codeinput.com). In that context, you get lots of events and payloads from Git/GitHub that typically require very little resources to respond to or relay.
The worker model made sense, so I developed the whole app around it. Now of course, knowing what I know today, I might have considered different options. But at the time, I read the description (and had some Cloudflare Workers experience) and thought this looked good.
Boolean is rarely enough for real production workloads. You need a 'processing' state to handle visibility timeouts and prevent double-execution, especially if tasks take more than a few milliseconds. I also find it crucial to distinguish between 'retrying' for transient errors and 'failed' for dead letters. Saving a few bytes on the index isn't worth losing that observability.
Iran is a democratic republic just like the 'democratic peoples republic' of North Korea is, i.e. not at all. It is remarkable how often entities which use the term 'democratic' do not live up to the concept it refers to
AFAIK the signature mechanism hasn't been defeated, so the attacker can only load software signed by the factory keys.
Which includes old, vulnerable versions and all patched, newer versions. By burning in the minimum version, the old code now refuses to boot before it can be exploited.
This is standard practice for low-level bootloader attacks against things like consoles and some other phone brands.
The problem with centralized social media is that the admins have power over you. They can ban your account with no recourse, censor some of your posts (or some posts you want to read), or even post something from your own account that you don't approve of.
Mastodon doesn't change this, it just changes who the admins are. It lets a person under the jurisdiction of admin A interact with a person under the jurisdiction of admin B, which is better than fully-centralized X, but it doesn't solve the fundamental problem. Your instance admin can still ban you with no recourse (account migration is incomplete, requires cooperation on both sides, and mostly exists to shut up Activitypub opponents who point these problems out). They're still just as (if not vulnerable) to government pressure as centralized social media, and considering that a single lawsuit could probably bankrupt most instances, I suspect they'd fold very very quickly. They can (and very often do) defederate from instances that post "too much nazi content", and if you disagree with the decision, there's again no recourse (you can migrate, but you won't get your lost relationships back).
> less fortunate people in corporate hellholes must face daily
I'm sorry but how tone-deaf can someone be? Over 12.000 people have been killed in the protests with some reports going up to 30.000 since then and here you are happy about the fact that Google cannot profile them anymore. Protesters are beeing shot on-masse in the streets and families from outside the country have no ideas if their brothers and sisters are even still alive. Have some decency.
