thimabi

8 ms·

91.
by thimabi 1y ago
If companies offered any sort of counter, users would quickly understand their usage patterns and adapt accordingly. But users are kept in the dark on purpose, often not knowing how many tokens they use, and sometimes not even knowing what
92.
Is Google about to destroy the web?(bbc.com)
4 points by thimabi 1y ago|1 comments
93.
by thimabi 1y ago
That’s just typical for any company that promises unlimited usage and then can’t deliver. The sad truth is that the “unlimited” wording attracts tons of users who will never come close to being a power user making a dent in their profits.
94.
by thimabi 1y ago
Things would be much easier if, instead of “unlimited”, companies simply offered great usage limits. Yes, hoarders would still exist, but their impact would be negligible. Even better: provide a counter displaying both remaining usage avail
95.
by thimabi 1y ago
I got in this thread with the expectation of seeing a comment like yours, so thank you for that. I think what worries the U.S. more is the likelihood of Pix spreading around the world. It’s such a great public program that I believe many co
96.
by thimabi 1y ago
I like 1Blocker. I’ve downloaded it when it was an inexpensive one-time purchase, and was grandfathered when it became subscription-based. Though I believe there remains an option to pay once and keep using it “forever”, now at a slightly h
97.
by thimabi 1y ago
> I can’t tell if the first sentence is sarcasm or not. Yep, it was. I wholly agree that the document feels less guided by the public interest rather than by various business interests. Yet that last goal is in a kind of weird spot. It f
98.
by thimabi 1y ago
I love how practically all goals in this Action Plan are directed towards incentivizing AI usage… except for the very last one, which specifically says to “Combat Synthetic Media in the Legal System”. Given that LLMs, for instance, are all
99.
by thimabi 1y ago
There are other comments joking about how this app will never secure funding unless it implements an AI feature… well, problem solved! Just have the app periodically flash AI-generated photos of women similar to the wife.
100.
by thimabi 1y ago
Thanks for the explanation! It’s always nice to learn something new :)
101.
by thimabi 1y ago
Maybe Tesla’s batteries have some sort of competitive advantage that Elon et al. don’t want to risk leaking to the rest of the world? (no idea if that’s the case, I have no specialized knowledge about batteries, just venturing a guess)
102.
by thimabi 1y ago
I find it amusing how we’ve come from treating AI as a novelty to developing a sense of how it writes in the space of a few months. That parent comment doesn’t even have the famed em dashes, for instance. Still, we are able to recognize it
103.
by thimabi 1y ago
I wonder if Apple itself offers repasting services via AppleCare. For someone like me, with little experience in handling electronics, it might be better than trying to fiddle with the MacBook’s internals.
104.
by thimabi 1y ago
> Do you believe in Perplexity long-term, or a) OpenAI moving into search, or b) Google moving their search into AI. I must confess that I’ve never used Perplexity, as ChatGPT works incredibly well for me. Never felt the need to try some
105.
by thimabi 1y ago
But note that Gemini via Workspace doesn’t have all the features of Gemini Pro, and it is notoriously behind regular Gemini in terms of feature adoption. So far, there’s no access to Gemini CLI (paid separately) and no way to selectively de
106.
by thimabi 1y ago
Right now, AI companies profit much more from wrappers, which pay API prices, instead of regular users, who pay loss-leader subscriptions. There’s little incentive for them to bother competing with wrappers today. Over time, however, I do t
107.
by thimabi 1y ago
However, as we’ve learned from the Epic vs Apple case, corporations don’t really care about bad behavior — as long as their ulterior motives don’t get caught.
108.
by thimabi 1y ago
Ask about the candidates’ previous experiences, setbacks, and lessons learned from them. LLMs can certainly make up all of that, but it is much harder for candidates to hide inaccuracies in their backgrounds, particularly over time.
109.
by thimabi 1y ago
Without it being attached to the AI hype, it surely is. In the current tech landscape, there’s a tendency to stuff AI into everything, so anything that doesn’t include it ends up being left behind.
110.
by thimabi 1y ago
Can you share some examples of things you’ve been building with AI? From your comments, I’d venture a guess that you see your AI-assisted work as a creative endeavor — an expression of your creativity. I certainly wouldn’t get my hopes up f
111.
by thimabi 1y ago
Do you think that, over time, knowing how the models behave, simply providing more/better context and instructions can fill this gap of wanting to provide taste and direction to the models’ outputs and actions? My experience is that, f
112.
by thimabi 1y ago
How do you type Cmd+C versus Ctrl+C? I find the Windows version more ergonomic because the keys are farther apart, so I press Ctrl with my left hand and C with the right one. Meanwhile, on MacOS, Cmd and C are closer to one another, so I en
113.
by thimabi 1y ago
Just delaying the y100k bug, I’d say
114.
by thimabi 1y ago
> Custom Search JSON API provides 100 search queries per day for free. If you need more, you may sign up for billing in the API Console. Additional requests cost $5 per 1000 queries, up to 10k queries per day. What I meant is that pricin
115.
by thimabi 1y ago
Can you recommend a specific piece of software for using an MCP integration for web searches with local LLMs? That’s the first time I’ve heard of this.
116.
by thimabi 1y ago
Unfortunately 100 queries per day is quite low for LLMs, which tend to average 5-10 searches per prompt in my experience. And paying for the search API doesn’t seem to be worth it compared to something like a ChatGPT subscription.
117.
by thimabi 1y ago
I’m thinking about building a pipeline to mass generate descriptions for the images in my photo collection, to facilitate search. Object recognition in local models is already pretty good, and perhaps I can pair it with models to recognize
118.
by thimabi 1y ago
Suppose I'd like to use models like this one to perform web searches. Is there anything available in the open-source world that would let me do that without much tinkering needed? I think it’s something that even Google should consider
119.
by thimabi 1y ago
From what I can tell, that was a bug and it has been fixed. I regularly use data with much more than 32k tokens in Gemini with my Workspace account, and context window issues are not a thing anymore. What does set Gemini via Workspace apart
120.
by thimabi 1y ago
Oh, you’re right, there are the legal issues as well. Just out of curiosity, do you see much difference in quality between the isolated model and the public-facing ones?

More ›