It already bugs me when ChatGPT describes how it is going to answer before answering, but it's 10x more annoying when I'm asking for a concise response without filler etc.
As an aside, I've noticed the self-description happens even more often when extended thinking mode is being used. My unverified intuition is that it references my custom instructions and memory more than once during the thinking process, as it then seems more primed than usual to mimic vocabulary from any saved text like that.
Right, it is currently incapable of providing a straight answer without clearing it's throat selling the answer. It reminds me of those recipe blogs that just can't get to the fucking recipe. It's bad writing! But it's not bad technically, in a style-guide kind of way.
Sometimes I wonder if the throat-clearing is an indispensable part of getting to the "good bits" that follow. Like, do those extra tokens give it more "room to think" even if they're basically meaningless in themselves?
The output tokens are the only information that is carried forward through each inference pass, so "more room to think" is incompatible with "basically meaningless". Perhaps one could imagine it somehow stenographically encoding information in its precise choice of meaningless throat clearing, but there are only so many variations on that theme - word choice is heavily constrained, so it doesn't feel like you could store a whole lot of information there without it starting to read froopiliciously.
I thought PP was saying that the "Thinking" text is only used for one turn, and the response text is the compressed thinking that survives into future turns.
1. Great if you have a wider screen (could never do it on my old 13" Macbook Air, for a 15" it's pretty good but for a 24" iMac it's perfect). But if you need the space youjust have it set to minimize by default, maximize on hover.
2. See the titles of your browser tabs, which is great when you are like me and never have fewer than 30 tabs open at once.
3. Easier to select browser tabs when you have many of them open (ie they don't get squished unreadably small)
I love Kagi as a search engine but the Orion UI feels too similar to Safari to really enjoy it as much.
I do enjoy vertical tabs, faster browsing, better privacy obviously. But "largely" is doing some heavy lifting in your mention of chrome extension support. I use about a dozen chrome extensions typically and about 4 of them are supported by Orion last I checked. Although of course #12 in Chrome is the Kagi search extension itself :)
The bookmarks bar seems consistently wonky though, with bookmarks showing the wrong logos (like Google Sheets showing up with the Google Docs logo, or ChatGPT showing some weirdly cropped version of itself), inability to rearrange bookmarks in a folder without opening the dedicated bookmark manager page.
If some basic usability things like this were fixed, along with adding tab groups (also big for me when I have 50 tabs open), I'd probably give it another go. Kagi search engine has largely replaced google search already for me so I'll definitely give it another go once these things are updated.
Please add a setting to disable tab hibernation. Opening my laptop on a plane only to discover most of the tabs have been offloaded makes using Orion impossible.
Ironically, as an American the only time in my life when I had a waiter effectively ask us (a couple) to hurry up so other diners could sit, was at a Michelin starred restaurant in Naples, Italy. We hadn't even been there an hour, weren't even done eating yet. Perplexes me to this day.
I won't get into the larger point your comment is making about power structures + definitions as I don't know enough about those histories to get into your assertions, but wanted to point out that the parent comment didn't seem to suggest these discrepancies were "biologically inert" as you were refuting (I'm also assuming you meant "inherent"). They were commenting on the a racial difference in educational outcomes. From my understanding that's largely a systemic issue, and regardless of shifting definitions/categorization, not a conversation about biology anymore.
That's interesting, it seems like it would accidentally penalize a lot of "good" posts too, like people asking questions to better understand a topic/perspective
This might not be entirely relevant since I'm not an expert in the 3d visual community but my friends launched this project and it might be something you'd be interested in
Thank you, and now I have a new pattern to watch out for in URLs. After all the energy we just had to go through to rescue all the goo.gl links when Big Evil decided that hosting a redirect webserver was too much trouble for them, and here we go again with share.google - crazypants
You can -- the real problem here is that each app could violate your privacy in different ways. Unless you break TLS and inspect all the traffic coming from an app (and, do this over time since the reality of what data is sent will change over time) then you don't really know what your apps are stealing from you. For sure, many apps are quite egregious in this regard while some are legitimately benign. But, do you as a user have a real way to know this authoritatively, and to keep up with changes in the ecosystem? My argument would be that even security researchers don't have time to really do a thorough job here, and users are forced to err on the side of caution.
What they do then is create an app where location is necessary, make that app spin up a localhost server, then add js to facebook and every site with a like button to phone that localhost and basically deanon everyone.
Probably not how you meant it but I chuckled.