coffeefirst

7 ms·

91.
by coffeefirst 5mo ago
That’s basically how I’m using it since I got wary about how the community plugins were being vetted. Core plugins and settings cover a lot. There’s one or two things I miss, but not enough to fork and review them myself so it’s clearly not
92.
by coffeefirst 5mo ago
Oh I know. In the past month I’ve moved several thousand dollars in spending away from companies that turned their support into a useless understaffed AI program. The disease has spread to six figure enterprise contracts hallucinating about
93.
by coffeefirst 5mo ago
In fact, I’m prepared to make no difference 80% faster and 50% cheaper.
94.
by coffeefirst 5mo ago
Maybe. But at least she gave a shit enough to actually write something you didn’t like. The sloplings don’t even bother.
95.
by coffeefirst 5mo ago
Ever notice how people making this claim never come with receipts?
96.
by coffeefirst 6mo ago
Yep. I’ve been using the Kagi Assistant with Kimi2 running out of providers in the US and Netherlands. The total monthly cost averages around $2. It be great if more people drop good open models. But either way stuff like this is the future
97.
by coffeefirst 6mo ago
Yes, that’s basically what I’m talking about. Without subsidies, that’s a lot of incentive to use more efficient/open weight models, but also to use them in ways that are less compute intensive—fewer tokens, shorter reasoning chains, l
98.
by coffeefirst 6mo ago
The other thing is as far as I can tell, the power-tool style modes where you use AI to boost focus and do quick research is both much cheaper and, by the time you account for all the damage control/prompt tuning/other externaliti
99.
by coffeefirst 6mo ago
I’ve thought about doing this. I mostly use YouTube for cooking and home maintenance. Give me the pasta making video shot in a commercial Italian kitchen with captions, and a grizzled old handyman demonstrating how to caulk the shower. Thes
100.
by coffeefirst 6mo ago
I really don't understand the people who use it for everything. It's become my first stop for search because it's doing it in bulk—read 50 results and lead me to something useful. But I just got Claude MCP connected to my per
101.
by coffeefirst 6mo ago
And this is the AI ethos: anti-contentiousness. Is it correct? Is it any good? Should I subject another person to this? Is it profoundly rude to not even read their email and just have a robot respond automatically? The slopmonger does not
102.
by coffeefirst 6mo ago
Yes. It’s been twenty years of watching people pick an fussy framework because they don’t want to learn the tool, then struggle to get the results they want because the framework is flawed but they didn’t learn enough to understand how and
103.
by coffeefirst 6mo ago
This is my theory too. There’s a predictable cycle where the models “get worse.” They probably don’t. A lot of people just take a while to really hit hard against the limitations. And once you get unlucky you can’t unsee it.
104.
by coffeefirst 6mo ago
It is legal to produce beer. It is legal to operate a bar. And we have a concept of a nuisance business that can result in fines or shutdown if your business constantly attracts egregiously bad behavior.
105.
by coffeefirst 6mo ago
There’s also a very real possibility that the skeptics are right.
106.
by coffeefirst 6mo ago
The voice is a real problem. I’m looking to do some basic home improvement stuff. 2 years ago I could YouTube how to do almost anything and find real people with plausible credentials demonstrating it. Today, first you need to dig through e
107.
by coffeefirst 6mo ago
Let’s also be clear: the asteroid doesn’t even need to be an energy crisis. If two money-losing companies decide that they would like to make money, the math gets ugly fast.
108.
by coffeefirst 6mo ago
This line jumps out to me: > Everything slows down. It’s like back in the old days when you really did one thing at a time. Why did we turn computers into frenetic, distracted multitasking machines? What would it take to reverse course?
109.
by coffeefirst 6mo ago
I have concluded the entire public discourse surrounding AI has no relationship to real stuff that you can go, test, and point at. There’s a loop of everyone is saying stuff because everyone else is saying stuff that turns into a sort of re
110.
by coffeefirst 6mo ago
I train up beginners pretty regularly and this is not a good analogy.
111.
by coffeefirst 6mo ago
Fine. Now give me back browser plugins that can actually do whatever I want them to. You can justify manifest v3 for security reasons, or you can can do this. You cannot do both without severe cognitive dissonance.
112.
by coffeefirst 6mo ago
Yeah… also, it’s just weird. Interfaces are important, they contain information and affordances, everything should not become a chatbot.
113.
by coffeefirst 6mo ago
That's how ChatGPT always seemed to me. One of the reasons I exclusively use other models is it would rather make something up than say "I can't find anything about that." But I'm starting to wonder about something.
114.
by coffeefirst 6mo ago
Definitely not. I really like Obsidian, but organize everything like a book, which gives just enough structure to know where everything goes without thinking about it, and no more. There’s just not enough there to make into a blog post.
115.
by coffeefirst 6mo ago
I think that’s correct. There seems to be a lot of fundamental limitations that have been “fixed” through a boatload of reinforcement learning. But that doesn’t make them go away, it just makes them less glaring.
116.
by coffeefirst 6mo ago
Eh, there’s tradeoffs. They’re real. But I’ve done plenty of this on teams back in the day before all these frameworks and it can absolutely work. It may even be easier now with JS modules.
117.
by coffeefirst 6mo ago
Maybe. Maybe not. But if this party is sustainable, what’s the deal with the rate limits that everyone is going on about?
118.
by coffeefirst 6mo ago
You don’t need anything this elaborate. Set parental controls on set up, pass a single flag to websites and apps, similar to the Global Privacy Control. No privacy is lost. Control is handed to the device owner, and implementation is techni
119.
by coffeefirst 7mo ago
Yep. I hate to predict the future but I’m betting on small, open models, used as tools here and there. Which is great, you can get 90% of the speed up with 5-10% of the cost once you account for the full costs in both time and money. The ec
120.
by coffeefirst 7mo ago
Wordpress, like SQL, is probably immortal. But it needs a better headless capability. Most separate front ends appear to be grafted on relying on plugins. Which doesn’t make a whole lot of sense.

More ›