ffsm8

Sincerely , I think you're blaming the AI incorrectly there. You just got incompetents on your payroll.

This in the same week that a women murdering her 3 kids is being called innocent by thousands of women on the Internet.

Very "interesting" times, indeed.

From the pitch itself

> Weedout is a $1.99 Safari extension for macOS that removes videos YouTube labels “Made with AI”

So no, none of that. Blows my mind if someone pays for that

Meh, almost nobody will want to live in most of Europe post AMOC collapse in 20-40 yrs

Look at the excerpt. They're not overwriting the sudo binary. The attack vector is real for malware running on a administrator user session which can be escalated to root via sudo.

It's a niche, but it's real. Esp. if you're targeting npm installed user scripts or similar

> What's so wrong if something is done to benefit the regular people once in a while?

Usually, this is money taken from people that aren't well off either... So it's not a clear cut positive.

Wherever your society decides to do such is obviously not for me to judge.

As a single almost 40yo person myself I'm significantly annoyed by all the money the state takes from my paycheck to sponsor people who quiet literally would spit on me and hold me in contempt. And while i ain't poor, my netto wage is lower then some of those people get From the state. But I'm not French, so no opinion on this particularly case.

Aka vibecoding platform.

A software that's optimized for agentic development, varying levels of review necessary depending on the technical choices of the factory developer. Some review may be automated via agents itself etc.

Ymmv obviously. Its definitely one of the funnest toy projects I've been working on for the last 5-7 months now (such a factory, mine is unreleased however - just for myself as of today)

> This, like almost every typing training system out there, suffers from my pet peeve

Did you know about

https://zty.pe/

It's a pretty old one at this point, but fantastic implementation

Fwiw, you may be misunderstanding them.

In large corps you usually have separate teams for the UX design to the frontend devs and the people specifying the features.

In such a scenario the UX designer can utilize such a bookmarklet to get a quick copy of the existing dashboard into figma to then iterate on the UX and try out different approaches for that specified feature.

This scenario would not be in violation of any terms and it's unclear what noobplus was actually thinking of

Mcp predates skills - and has a more granular permission model then skills + bash commands.

You're pitching a solution that's incredible brittle and unnecessarily complicated if you think about it in technical terms.

For your feature to work you need bi-directional communication between the otel receiver and your application - that's still doable in general, but now you want a synchronous "upgrade" to traces.

Now we're talking about a massive performance impact - and you need to somehow cache all otel data locally so they're available for the upgrade and only then submit then.

It is a architecture that's not very smart, honestly. And precisely the reason why you'd simply submit everything and let the receiver figure out which samples it wants to keep - as thorian pointed out earlier.

"caveman speak" skill, need I say more?

People aren't particularly bright. That's why the scientific method was developed to counteract our built-in tendency for... Unorthodox approaches

Thats a very spicy threat model post the age of IoT.

Any stray device you may not even be aware of can rip that assurance apart.

A fully locked down network can in theory enable this however, very few enterprises actually are strict enough for this to be enforced across the whole network.

It's also unclear wherever the subscription price is the real cost, or the API.

I suspect it's closer to the sub price and anthropic is just milking their API users, but that's something you'd only know from the inside

It is the reason why user stories were established, so that the person interpreting the spec knows why. Generally speaking this reduces the likelihood of this gap occuring with humans significantly. At least when they're not incompetent.

Right, I was wrong. The best calorie/$ ratio is with a bottle of cooking oil. (7-9k calories/$)

But you cannot eat that, so potato chips are a distant second (at least from the products I checked)

And no, the calories in chips added by the oil offset the lesser amount of weight you buy per unit - at least in my locality.

If I just look at calories per dollar, potato chips seem the most cost effective - at least as long as you take the cheapest ones

Pure potatoes are somewhere around 200-500 calories per dollar, cheap potato chips are around 1000/dollar

Pure noodles or rice might be even better, I didn't check those admittedly.

If you're going for calorie/$ then nothing beats potato chips

at that point, i'd put bash script up as being moderately more secure as its generally easier to audit them vs a binary.

but i'Ve also gotta say that random binary download over web is also incredibly rare - usually its either a combination of both (the curl|bash ending in a random binary being downloaded) or the user actually installing via a packagemanager like apt, zypper, yum, dnf etc - and those packagemanagers generally do audit the main repositories, so theyre basically as secure as you can get in those contexts.

but of course, everyone has their own thread model and i dont work in security (●'◡'●)

in case anyone else is mildly curious (i didnt know that was a thing)

> The reason it exists is shell flexibility. Traditionally root's shell is kept as a statically-linked shell like /bin/csh or /bin/sh so that the superuser can always log in even in single-user mode or if dynamically-linked shells in /usr/local break. The toor account lets an admin have a UID 0 login with a fancier daily-driver shell (bash, zsh, etc.) without touching root's safe configuration.

you can detect `curl | bash` server-side and serve a different payload for those (compared to curl -O file, wget etc), hence its an effectively undetectable attack vector.

Executables on the other hand can be inspected and prodded, so the likelihood of something going amiss and consequently security agencies finding out about it is significantly higher.

neither of those is secure of course, we're just discussing different levels of dangers. And curl|bash being worse, albeit not that much

(and the -L here is the extra cherry on top. piping a redirect to a shell is just monkas)

According to it's IPO documents, SpaceX is a 99% ai company, with a tiny offshoot doing space shipping, essentially

also, session cookies for eg login and error tracing etc dont actually need a banner. you only need to use it if youre participating in the internet stalking behaviour by integrating eg google analytics and similar integrations

Frankly, I would've agreed with your opinion... Maybe 20-30 years ago?

I think it's clear that nothing will be done - and initiatives that end at a national level are meaningless from a climate perspective.

You should definitely still try to get your locality to do them, because they increase your quality of life by reducing local pollution... It climate change relevant? No.

15-20 years ago I would've said we need a more global approach to this, actual binding agreements - enforced globally through a body that has the ability to force the nations to actually adhere to them.

But now, in the last 5-10 years I've most certainly lost any kind of hope that such a initiate would not end being leveraged by someone like trump to sabotage others.

Honestly, I currently consider any climate change oriented political decisions to be irrelevant at best. It's gonna happen, so if you do something - do it for your local environment. If you try to signal with global warming I am forced to see you as at best a hopelessly naive person, as I was 30 years ago

Youre the one that made the intellectually dishonest argument that scale is about cost and efficiency while the discussion was about scale of the surveillance.

from your point, nothing is cheaper and more efficient then mandating an app on all phones through which the governments can track everyone, and consequently whoever doesnt have it would be a criminal. it completely tracks from your argument. You may need to make the app stop tracking while theyre in the home, but thats a technicality. voila. no longer an issue according to your point

I mean I don't have anything against go, but frankly - it's not really better then modern Java.

Each have their trade-offs, both can support native compiled application code. Some architectures are easier to review and code in golang, but others go much better with Javas richer ecosystem and better composability.

Read the article to find out.

Token use didn't seem to be a criteria from my casual reading, but maybe you can illuminate me what section pointed to that, I may have been too superficial in my reading

its vibecoded. just point another vibe based harness at it and youll have it in an hour or two. this thing is as trivial as you can get get, scope wise

Reading your quote really gets me wondering who the people making those kinds of analysis are...

It's like they haven't maintained any actual software, because the criteria they choose is... Completely irrelevant?

The things that matter are tooling, orchestration and ecosystem - as well as how the LLM will actually implement the solution for a task

LLMs constantly do idiotic things. If you have good libraries to utilize, the likelihood of the solution actually working goes up because they no longer need to implement the hard part.

If you have orchestration for dependency injection, code generation, meta analysis etc

Tooling like the way otel tracing is integrated, openapi generation etc is also invaluable because every time the LLM does something the likelihood of it being hallucinated/wrong increases etc

You'd need to implement a nontrivial project in different languages, then add nontrivial features across them and only then start by rating eg correctness and incident occurrence after the final output

But token use on a one shot? Completely irrelevant as far as I see it.

I understood this comment chain to be about the technology archetype of smart glasses with cameras in them, not about Meta glasses.

If it was about the latter then I'd agree that there is no usecase for them

I think eg police and military personnel should be required to wear cameras while on duty and interacting with people (So not during their whole shift, but whenever they're knowingly entering an interaction)

This may be a body camera or a camera built into glasses. That is a very legit usecase as far as I'm concerned.

There are other professions in which such a device would be a net benefit, that was just a simple example.