Except there already was "something" that the creators of MCP just ignored.
Imagine a world where you could:
* Configure your favorite harness/chat client with any OpenAPI spec for an API that supports Oauth2.
* The harness would walk you through the Oauth flow and securely store the token.
* And then insert some tools for discovering the API methods and making requests in to the context.
* The agent could then formulate a request, call the request tool, and the harness would 1) makes sure it's allowed to make a request to that API, and 2) insert the Auth Token into the request.
It would be basically exactly the same way MCP is setup today, except all you would need is an OpenAPI spec. You wouldn't have setup a server for a janky new standard that's half implemented slightly differently by every harness/chat client.
It sure does feel that way. But I've often maintained that Apple took a look at how people were using features like this on other operating systems and said, "NO! We're gonna 'think different'!" And no one liked their way of "thinking different". But they keep doubling down, believing that once we see the genius, we'll all come around.
I haven't yet come around. Unix had it figured out back in CDE days. i3 was my happy place... now it's Niri. As long as I can get to my workspace with a key combo THAT I DEFINE, I'm probably okay.
I don't know -- my conspiracy theory for many years has been that Apple isn't really very interested in multiple desktops at all, and there's just this one engineer who's made it his life's mission to bring the idea to the masses. But the masses also aren't really very interested in multiple desktops. So every two or three versions, he screws around with the UX and says "surely this time everyone will see the genius!" And they do not.
Personally, I almost never use more than two virtual desktops on macOS, which means most of the changes and bad decisions sail over my head. I don't think I really seriously used more than two virtual desktops on any OS later than BeOS, though -- which IIRC did it really well! -- which I suspect is less a function of macOS doing it badly than a simple function of rarely feeling like I'm cramped on a 27" 5K monitor. (Yes, I know half of you are like "I have three 24-inch monitors and each one needs nine virtual desktops to accommodate all of my windows." That's fine. You do you.)
I mean... I have a 34" widescreen, and yet, I spend most of my time on my laptop screen. So having multiple workspaces is pretty helpful for focus. I think your theory is a sound one. Once they realized their main business is a single screen with ONE APP focused at a time (yes I know, iPad can do more, Duo will also do more), they stopped caring about anything else.
I just want them to make the animation faster. It makes Spaces completely useless to me that it takes what feels like 2 seconds to shift focus to a window on the new space when I swipe on my trackpad. Worse, it takes longer the higher refresh rate your monitor is; it's not too bad on 60Hz screens.
You can turn the animation off - but, when you do, there's still a ~2s pause when the workspace you switch to is dimmed and unresponsive so it feels just as janky.
There's simply no way to quickly flick from one space to another and then back again, which is something that could be done more or less instantly on Linux desktops back in the late 90s!
You can't turn the animation off, you can change it to be a fade animation instead of a slide animation. As you say, it takes the same time regardless.
I'm not sure it's strictly red state/blue state. I've live in Massachusetts and Minnesota for a while, where they made it very easy to vote so that tracks.
However, now I live in New York, where they seem to go out of their way to make it difficult. The registration cutoff is several weeks before the election and for at least one election, I found myself wrongfully expunged from the voter roles.
I think it may just be how corrupt the state government is.
Kroger's delivery service wasn't that bad at all, it was fantastic actually.
Of course, some things are too good to be true, they apparently lost hundreds of millions of dollars on the effort before closing it down earlier this year. At least they did in my area in January, my understanding is they ended it everywhere they didn't have a physical store presence.
I buy "pickup" orders online all the time through the store's app where you can also buy instacart delivery. It's still shopping online but there's no additional cost nor tip for ordering pickup.
Is that assuming Instacart/DoorDash? Click and collect or even delivery from grocery stores isn't nearly that bad, right? I thought it was pretty competitive.
If you order at least once a month it is probably worth upgrading from Walmart+ to Walmart+ InHome. That's $40/year but it has these advantages:
• Delivery is by a background checked salaried Walmart employee in a company van rather than third party contractors.
• Tipping is not allowed.
• Instead of just delivering to your doorstop you can have them deliver to your garage, or even inside your home where they will even unpack them and put the perishable items in your fridge and leave the rest neatly arranged on your counter top or table.
Since the minimum order for free delivery with Walmart+ is $35, a 10% tip on a single order would be more than the $3.33 that a $40 year annual subscription divide by 12 is (and most people would tip $5 on a minimum order rather than 10%).
Very interesting. The person I know that thoroughly uses the services avoids buying some perishables (milk, meat) because they get to unsafe temps and quickly spoil. Maybe this would solve their problems!
For a lot of people their time and their effort is worth more to them than their money unless it is something they particularly enjoy doing themselves.
In my experience, the docs for something like setting up a dev env are typically greatly improved by the second person who sets up the dev env, not the personal who originally wrote the docs.
In that case, when the docs are not associated with a code change, you want to make getting those improvements into the docs as frictionless as possible, otherwise the changes aren't going to get made.
Personally, I've found that making docs updates incredibly fast + easy to be far more valuable than anything you get from forcing doc changes through the full SDLC process. If someone has feedback on your docs changes they would have shared in a review, they can just update the docs instead.
> In my experience, the docs for something like setting up a dev env are typically greatly improved by the second person who sets up the dev env, not the personal who originally wrote the docs.
In my experience, this is also true of a lot of code as well. Your dev scripts should probably have much more relaxed standards than your service source or CI/CD. Ideally I could define merge requirements by directory without doing some weird shenanigans with the CODEOWNERS file and a bot.
World's best AI models doesn't offset public resentment caused by putting data centers in residential areas, openly buying off the White House and Congress with PAC "donations" and constantly threatening to eliminate everyone's job.
So Mamdani or Sanders putting out videos on tiktok to engage with young progressives, that's also a "PR campaign"? Or is "PR campaign" one of the phrases where you only use against other people, eg. "we're doing voter outreach, they're doing PR campaign"?
They're buying the credibility of the influencers.
Influencers have the greatest influence on their audience, especially repeat viewers who follow or subscribe to them. You need to establish credibility with an audience before you're worth anything to the marketers. This takes time to build.
OpenAI's hypothetical robot influencers have no audience.
Not sure if I'm missing some sarcasm in your post - apologies if so. But as always, if it actually worked, they'd just be making money off of their many million-follower accounts, instead of telling you how to do it.
This is the mistake they made. Anthropic and OpenAI are great technically. I love them. But they are absolute crap at making most people feeling like its relevant to them or that AI companies care about them. So its' really easy to be dismissive of the tech and oppose things like data centers. I think the influencer route is a quick fix. They need to appeal to everyday problems but, understandably, the founders like Dario and Altman think very high level. They think they are doing what's best in a way, but not in a way that's relateable to most people. LLMs need to be more human friendly; example, solve little problems like actually writing without sounding like AI
I think the OP was referring not to their public image, but to the incongruity of paying human influencers instead of deploying computer-generated influencers created by their world-beating models supposedly worth $30T.
OpenAI is trying to level up their growth marketing to match Anthropic’s. Tactics in growth marketing (speaking from my experience leading in this space) are not public knowledge as it’s a cat-and-mouse game on multiple fronts (ie CAC goes up and changes the economics when a high ROI source is found, etc. etc. etc.), which makes hiring the right person vital as information asymmetries are where the ROI is (ie more revenue per dollar spent on marketing). Anthropic hired the right person for strategy and execution long ago in LLM-years, and there is a good podcast on the story here:
https://www.lennysnewsletter.com/p/anthropics-1b-to-19b-grow...
And that is incredibly unfortunate to their business strategy.
But... food for thought: people are good at detecting slop, because it looks like slop. If it does not, people won't detect it. So there is a confirmation bias, where the "obvious slop" looks obvious, but the "better slop" goes by undetected.
1. I would argue that human directed and edited "better slop" is just... average content. Which leads to my next point.
2. LLMs have raised the floor of digital content creation ("democratized content creation"). Before lots of things would have not been created due to lack of knowledge, interest, motivation. Now the baseline for making something is a 10 word prompt. Personally I think it's a huge loss. Yes, we will get 5% "diamonds in the rough" who would have never approached a field due to gatekeeping, discrimination, whatever. Will that outweigh the other 95% of lazy creators flooding every field with around average content? I'm not convinced. There is a cost to that flood.
I think the point of those comments is this could have been much simpler for everyone if the model providers had given us some tools around existing standards.
Instead they invented something new/weird/complex standard, and then every client implemented slightly differently (and different parts of it).
Imagine a world where you could:
* Configure your favorite harness/chat client with any OpenAPI spec for an API that supports Oauth2.
* The harness would walk you through the Oauth flow and securely store the token.
* And then insert some tools for discovering the API methods and making requests in to the context.
* The agent could then formulate a request, call the request tool, and the harness would 1) makes sure it's allowed to make a request to that API, and 2) insert the Auth Token into the request.
It would be basically exactly the same way MCP is setup today, except all you would need is an OpenAPI spec. You wouldn't have setup a server for a janky new standard that's half implemented slightly differently by every harness/chat client.
reply