It really depends his good it is. There are so many shows where stories just don't pay off well. I could see an AI developing a show with a complex web of throughlines kept logically consistent and fleshed out.
But why? I feel like there is always this implied notion that media is just about “being good.” But “being good” requires a lot of things, and a big part of that is somebody sitting down and deciding they have a story to tell. An LLM cannot do that. It cannot speak from experience, it cannot explore novel concepts, it cannot express actual meaning.
A pretty photo is not enough on its own. Hotel lobbies have all kinds of pretty photography. That’s not the stuff people hang on the walls in their home. There is nothing wrong with it, it is fine to have hotel lobby photos. But acting like that is the ultimate target, that we should use AI to just manufacture this at a large scale… I mean what is even the purpose of that? Am I missing something here? Do people really just want to sit around with their local agent generating facsimiles of existing shows/movies for cheap entertainment?
There is always space for cotton candy and general purpose work, but we should not be advocating for a massive facsimile machine to drown out all media. People have been whining about Hollywood not writing new things for decades, so the alternative is to accelerate that? I just don’t understand the draw. I have no desire to sit around watching my computer make a fake, derivative Star Wars show for me to watch later.
MCP is a tool more for security than anything else. If you give your agents access to an API key, there's a chance they can accidentally or maliciously leak that key. If the MCP server has access to the keys instead, it takes that possibility away. That's not always something you need to care about, but it is very important for some people's threat model.
> MCP is a tool more for security than anything else.
+1 to this; API keys mean:
a) Your agents are overprivileged by default
b) Your API key is more likely to get into logs and pre-training / leaked as FrinkleFrankie mentioned
c) You can't manage it centrally with granular tool policies in Enterprises. (e.g. "you're not allowed to read slack channel for #finances")
Also, you can always collapse MCP to CLI, so MCP as auth/security middleware makes a lot of sense.
> MCP is a tool more for security than anything else.
The agent can get to the resource through the MCP server or using API key. I personally do not see the benefit MCP is providing here. Sure you can reduce the exposed surface at MCP layer, but I do that at the API layer. I don't need to add another layer here.
I can kinda understand if you do not have control of the API layer and/or you have to expose the API layer to the public Internet as well. Most of the time that is not the case for me.
There are many ways to do this without using mcp, for example one of the many proxy servers that inject the token into requests. I use a custom solution that redacts all secrets to agent transcripts before the agent sees them in a hook.
I find it way better to be able to confidently tell agents to use CLIs than worrying about partially implemented mcps that need configuration and are often yolo’d with npx @latest anyway
My own approach is to put the API key in the MCP server, apply principle of least privilege there, and firewall access to MCP with agents being on same private network with Tailscale.
Not only giving an API key to an agent can leak to the model because of harness issues or too broad reading rights, but also most providers don't give the ability to apply principle of least privilege to an API key.
I don't want to give an agent full R/W access to any of my services/accounts.
The agent doesn't, the harness does. It's separate from a normal conversation / agent runtime environment. How do you suspect auth keys can leak from an MCP that's been added to chatgpt.com?
With a different key, obviously. This lets you keep the MCP server firewalled to your local network only while still allowing home assistant itself to access the internet
I'd rather use the fancy models while absurd valuations are subsidising them and then pick up a couple grand worth of their old hardware in a few years when they're eol.
I've had a ZFS array running since ~2009. It's gone through upgrades and disk replacements, etc. Have not suffered a data loss in that time. ZFS I'd the goat.
I'd say anyone who chooses to kill a civilian, whether based on an order or not, is a war criminal. I really don't care if they thought they were there for a noble cause or not.
What if they choose to kill a person who was a civilian until they were drafted into the enemy army the day before?
What if the civilian they kill has a substantial effect on the enemy's military capabilities, for example by being someone famous who loudly endorses the honour of signing up to fight for one's country, thereby motivating thousands of young men to do just that?
What if killing a soldier’s children might make him lose his resolve or act erratically? What if destroying an entire city full of millions of civilians might make the enemy lose hope and surrender? What if poisoning the land and water for decades could prevent the enemy from reconstituting itself? What if killing all the women could prevent the production of future soldiers?
Once you open the door, where do you draw the line?
I interpret your examples as further evidence (in addition to the examples I gave) that there is no sensible justification for distinguishing military from civilian casualties. Was that your intention?
On the contrary, I was fully intending to show that you need a hard line between civilian and military when considering military action, or mass atrocities and even genocide are almost certain. It’s a little concerning that you didn’t understand that, or willfully misunderstood that. Make sure you don’t lose your humanity.
I think someone who considers a civilian's life to be starkly more valuable than the life of a person who was a civilian yesterday but, through no choice of their own, is a soldier today, should be mindful of their own humanity.
It is precisely the concept of such a "hard line" -- which by design enables such inhumanity to be passed off as inconsequential -- that must be brought out into the light and examined.
Windows 2000 was incredible. It was so sleek, fast and worked so much better than XP. I'm sure it still required a lot of tinkering compared to a modern OS, but I still miss it.
reply