I guess it's great if you don't give a shit about code quality, particularly in a larger project.
What I see are three tiny little projects that do one thing.
That is boring. We already know the LLMs are good at that.
Let's see it YOLO into a larger codebase with protocols and a growing feature set without making a complete mess of things.
So far CC has been great for letting me punch above my weight but the few times I let it run unattended it has gone against conventions clearly established in AGENTS.md and I wasn't there to keep it on the straight and narrow. So a bunch more time had to be spent untangling the mess it created.
Yeah I don't know if this is a skill issue on my part, the nature of my projects, the limits of Sonnet vs. Opus, or a combination of all of the above, but my experiences track with all of yours.
From the article:
> The default mode requires you to pay constant attention to it, tracking everything it does and actively approving changes and actions every few steps.
I've never seen a YOLO run that doesn't require me to pay constant attention to it. Within a few minutes, Claude will have written bizarre abstractions, dangerous delegations of responsibility, and overall the smelliest code you'll see outside of a coding bootcamp. And god help you if you have both client and server code within the same repo. In general Claude seems to think that it's fine to wreak havoc in existing code, for the purpose of solving whatever problem is immediately at hand.
Claude has been very helpful to me, but only with constant guidance. Believe me, I would very much like to YOLO my problems away without any form of supervision. But so far, the only useful info I've received is to 1) only use it for side projects/one-off tools, and 2) make sure to run it in a sandbox. It would be far more useful to get an explanation for how to craft a CLAUDE.md (or, more generally, get the right prompt) that results in successful YOLO runs.
Watching my website's firewall and ssh logs show all the various hacking attempts is calming in the same way that watching waves crash on to the shore is.
Back in the day, I made the mistake of hooking up a fresh Windows XP (at least I think it was; pre-SP2) install directly to the internet. There was no firewall or NAT to protect me. The machine got pwned almost immediately.
Just yesterday I saw a Sora-generated video that purported to be someone filming a failed HIMARS missile failing and falling on stopped traffic and exploding on the 5 in Camp Pendleton on Saturday. (IRL they were doing some kind of live-fire drill and it did actually involve projectiles flying over the freeway.)
While there were some debris instances IRL the freeway was completely shut down per the governors orders and nobody was harmed. (Had he not done this, that same debris may have hit motorists, so this was a good call on his part)
You could see the "Sora" watermark in the video, but it was still popular enough to make it in my reels feed that is normally always a different kind of content.
In this case whoever made that was sloppy enough to use a turnkey service like Sora. I can easily generate videos suitable for reels using my GPU and those programs don't (visibly) watermark.
We are in for dark times. Who knows how many AI-generated propaganda videos are slipping under the radar because the operator is actually half-skilled.
> I can easily generate videos suitable for reels using my GPU and those programs don't (visibly) watermark.
Curious what you used. I have an RTX 5090 and I've tried using some local video generators and the results are absolute garbage unless I'm asking for something extremely simple and uncreative like "woman dancing in a field".
I am pretty sure what you want is doable on a 5090 with some effort but it will not be just a text prompt to video. More like input key frames as images and interpolate video between them.
I don’t give a fuck how cheap you make bread and circuses. The only goods and services I give a fuck about anymore are the ones that won’t be made cheaper with automation.
Speak for yourself. I want cheaper building costs so we can build more housing, cheaper and safer vehicles, higher quality food so we can all be healthier, better medical technology and medicines so we can solve more diseases, and new washing-machine-like technologies so I can spend more time with friends/family. That’s not to mention that greater leverage on my labour would give me even more flexibility to choose the work I want to do, and how much I want to work.
Bread is already so cheap as to not notice the price most of the time. But other goods and services are absolutely not that cheap. And there’s certainly higher quality that could be achieved, especially in areas like medicine. It is a lack of imagination to not see all the ways in which cheaper goods and services could improve our lives.
Making all of those things cheaper is great, as long the automation isn't also making everyone poorer at an equal or faster rate. It doesn't really help if house prices and food prices are cut in half if most people lose their employment because of automation.
I think the concern is that true human+ AGI and advanced robotics would obsolete so many roles that it doesn't matter if things can be made more efficiently, because nobody will have any money at all. If/when AI can do my job better than me, it isn't giving me leverage, it is removing all leverage I have as someone who puts food on the table through labor.
In the interim period before that happens then sure, the automation is great for some people who can best leverage it.
On the path to “AGI” I would expect a lot of short-term pain as people lose their jobs while unemployment is still around normal levels. But if unemployment rises too much, we would pass laws to protect people, like greater corporate taxes to fund things like UBI.
But honestly, if we have this level of automation it feels like it would be very hard to predict how society will evolve. I would expect our current model of work-to-live to become untenable, and we’d move to something else. I doubt that transition will be easy.
It's never going to go down like that as long as companies are required to serve shareholder interests above customers' or employees'.
Instead all these automation tools are and will be used to cut corners and optimize on cost. Quality, peace-of-mind, and increased free time will be the sales pitch used to placate us plebes. But we all know what the executive dipshits will really care about.
Most people here could choose to work less than full-time hours if they wanted to. I already do (although I do it so I can work more on my own projects, to be fair).
Although, maybe going against the hedonic treadmill is against our nature. There’s always a nicer house in a better neighbourhood to work for. But I at least want more people to have the choice to work fewer hours through higher wages. That might not come for free with economic growth, but it certainly won’t come without it.
I've found it to be excellent but 2.5 seems to experience context collapse around 50k tokens or so. At least that is my findings when using it heavily with Roo Code
I've since switched to Claude Code and I no longer have to spend nearly as much time managing context and scope.
I don't want to take the time writing up a cogent response to an article someone didn't bother taking the time to write. With this particular article, there were a couple of points I wanted to respond to, before I realized there was no human mind behind them.
I've always liked the HN community because it facilitates an intelligent exchange of ideas. I've learned a lot trawling the comments on this site. I don't want to see the energy of human discourse being sucked up and wasted on the output of ChatGPT. Aggressively flagging this stuff is a sort of immune response for the community.
What I see are three tiny little projects that do one thing.
That is boring. We already know the LLMs are good at that.
Let's see it YOLO into a larger codebase with protocols and a growing feature set without making a complete mess of things.
So far CC has been great for letting me punch above my weight but the few times I let it run unattended it has gone against conventions clearly established in AGENTS.md and I wasn't there to keep it on the straight and narrow. So a bunch more time had to be spent untangling the mess it created.
reply