snappyl's recent activity
-
Comment on Google's new AI video tool floods internet with real-looking clips in ~tech
-
Comment on Google's new AI video tool floods internet with real-looking clips in ~tech
snappyl Link ParentI was actually just thinking about this myself not long after Google I/O. I apologize; I'm going to jynx this but what I came up with is what if Google updated the YouTube ad platform? Brands...There will be commercials completely built with AI very soon. The ones that do not really need in depth acting, but a person walking by or showing off their shoes.
I was actually just thinking about this myself not long after Google I/O. I apologize; I'm going to jynx this but what I came up with is what if Google updated the YouTube ad platform? Brands could pay for a targeted campaign with maybe a general theme and then just let Google run with it. Google could take that, plus their user targeting/personalization infrastructure and dump some kind of "make an ad script" prompt into Gemini for a user or groups of users. Then take that output and dump it into Veo. Then take that output and put it into your YouTube videos, per their last "put ads into the best locations" development. Boom. Hyper-personalized ads at the perfect time for maximum click-through..
Cost-wise I don't know how this pans out, but I would be shocked (SHOCKED!) if this wasn't being worked on.
-
Comment on Will the humanities survive artificial intelligence? in ~tech
snappyl Link ParentIn my experimentation I've also noticed that different chatbots have what I'll call different personalities. I said in another comment that I just got a ChatGPT sub and have been comparing it to...In my experimentation I've also noticed that different chatbots have what I'll call different personalities. I said in another comment that I just got a ChatGPT sub and have been comparing it to my Gemini usage. What I've noticed with my style of prompting is that ChatGPT tends to hallucinate more and speak with greater confidence on topics it "knows" less about, where Gemini tends to hallucinate much less (so much so I fear I'm at a point where I tend to trust its output on its face) and it'll use less certain language when its confidence is lower.
-
Comment on Will the humanities survive artificial intelligence? in ~tech
snappyl Link ParentI've been having an on-going research project on the benefits and harms of AI, using Gemini predominantly. This month I got a ChatGPT sub too, so I could compare the two and I've noticed that the...The professor seems very impressed by the results of trying AI for a class. These sound like better results than anything I got out of it, but perhaps I’m doing it wrong:
I've been having an on-going research project on the benefits and harms of AI, using Gemini predominantly. This month I got a ChatGPT sub too, so I could compare the two and I've noticed that the responses I get from Gemini are better for me. For my use, they tend to be more accurate, the possible hallucinations tend to be couched in more uncertain language ("may", "could", etc), and the larger context window availability lets me have larger chat sessions without it suddenly forgetting what we're doing. I don't know if my greater success with Gemini is because it is better or if it's because, like a programming language, I've come to understand it? If you wanted to use some chatbot for something, it may be worth while to test drive a few to see if maybe you just don't "get along" with what you're currently using.
I don’t think the scattered archives of primary sources are plugged in quite yet, though?
Oh this is a topic I want to explore more deeply in my research next. For now I know that
- Google has a deal with reddit, so Gemini can use that information without restriction
- OpenAI has a deal with Axel Springer, so any of their publications are likely available for use here
- I've seen Gemini return results using information from Harvard Business School, ACM, JMIR, American Psychoanalytic Association, and other schools. ChatGPT has similar sources in my research, but school selection varies depending on how each bot searches.
- Both bots really like arxiv.org for my research.
It's important to note, I suppose, that when I'm asking both for output I do ask them to favor academic sources where possible. So that's why we see a lot of those here.
-
Comment on End of 10: Replace Windows 10 with Linux in ~tech
snappyl Link ParentI waffle between Linux, Windows, and MacOS all the time, but the SCOOBE is like the first thing I disable any time I decide to hop back onto Windows. If you are sufficiently annoyed and never want...It's annoying to have to engage with the OOBE in Windows, but that dialogue comes up maybe once a month and it's a few clicks to get rid of.
I waffle between Linux, Windows, and MacOS all the time, but the SCOOBE is like the first thing I disable any time I decide to hop back onto Windows. If you are sufficiently annoyed and never want to see it again, this guide should put you in a better place.
-
Comment on What's something that makes you feel like we're living in the future? in ~talk
snappyl Link ParentI am consistently blown away by my interactions with LLMs. I have a sort of ongoing personal research project into the benefits and harms of LLMs (currently Gemini is doing some digging for me...Same here for me. Despite all the negative parts of how they were created and how they are used nowadays, the technology is astonishing.
I am consistently blown away by my interactions with LLMs. I have a sort of ongoing personal research project into the benefits and harms of LLMs (currently Gemini is doing some digging for me regarding impacts in the legal system) and the overall gist on whether they are good from an output perspective is "it depends". For me personally, having access to an LLM is better than not.
Let me give you an example: I've been looking at getting out of my apartment and moving about a mile away from where I am. As part of that search, a friend suggested checking out crime rates to make sure the new neighborhood was an acceptable level of crimey to my current neighborhood. So I tossed that as a question into Gemini deep research and it gave me back a solid summary of the differences. I asked it about another neighborhood. Boom. Same deal. In each case, I did verify its output with official sources and it all lined up. So then I thought, hey how did this happen? These neighborhoods aren't exactly distant from one another, so I asked Gemini and it came up with a report that matches the vibes of what I understand from the area historically. It would have taken me a day to do what Gemini helped me do in an afternoon.
As far as how they're made? eeehhhhhh. I'm glad Google does the honorable thing here and avoids visiting websites that ask politely for Google to not visit (robots.txt) and annoyed that others don't. Does that make Google not evil? I don't know. Is OpenAI evil? Also don't know. Am I evil? Could be...
Anyway, looks like my report is done and LLMs in the legal profession are racist. So I'm going to go dig into that and see what's up! Should be fascinating!
-
Comment on Confess your food crimes in ~food
snappyl Link ParentI don't know if this is a recommendation as much as it's just a suggestion: Follow Your Heart appears to be coconut based, rather than cashew based. Daiya also appears coconut based, but between...I don't know if this is a recommendation as much as it's just a suggestion: Follow Your Heart appears to be coconut based, rather than cashew based. Daiya also appears coconut based, but between the two I would say FYH tastes more cheese-like and Daiya tastes more like what someone who has never had cheese imagines it tastes like based on scent alone. Miyokos uses oat and coconut milk it would seem. It's been a minute since I've had anything from them, but my main complaint with their cheese is it's crumbly and falls apart easily. I don't remember the flavor being bad/weird, though.
I'm currently using the FYH smoked gouda on sandwiches and I'd say it's good. My taste buds are all kinds of wrecked though due to my own weird food styles so, again, I can't say these are recommendations.
Food crime though? Spaghetti. Just standard spaghetti with wheat noodles and mushrooms in a tomato sauce. No meat but with tempeh. Perhaps it is but a misdemeanor when measured up against pizza au jus where the au jus is coffee.
-
Comment on <deleted topic> in ~tech
snappyl Link ParentEugen Rochko, head of Mastodon development, commented on that a few days ago actually. Looks like sign ups are up for Brazilians on MastodonEugen Rochko, head of Mastodon development, commented on that a few days ago actually. Looks like sign ups are up for Brazilians on Mastodon
-
Comment on US athletes are taking full advantage of free healthcare in Olympic village in ~sports
snappyl Link ParentI've had to call the IRS a couple times and each time they've been ludicrously helpful and accommodating. That said, there is an obvious need for more agents. My wait time was measured in hours....I've had to call the IRS a couple times and each time they've been ludicrously helpful and accommodating. That said, there is an obvious need for more agents. My wait time was measured in hours.
If you have to call the IRS, the one piece of advice I can offer is this: realize the agent isn't at fault for any of this and treat them like a normal person. If you're the one person not yelling at them that day, they will move heaven and earth to help you with your issue.
-
Comment on Apple Intelligence in ~tech
snappyl Link ParentFair and a good call out. I was assuming that at some point the font rendering was handed off to Windows and they could leverage that, but I also know about 0.0% on how that actually works.They could probably integrate with their own reasonably well, but win32, MFC, WPF, UWP, and WinUI
Fair and a good call out. I was assuming that at some point the font rendering was handed off to Windows and they could leverage that, but I also know about 0.0% on how that actually works.
-
Comment on Apple Intelligence in ~tech
snappyl Link ParentFirst I have to say that I don't actually know how these two systems work -- I only know "how they work" based on statements I've seen from the two companies, so this could be wrong. That said,...First I have to say that I don't actually know how these two systems work -- I only know "how they work" based on statements I've seen from the two companies, so this could be wrong. That said, from what I understand, I think there's actually a pretty big difference between Microsoft Recall and Apple Intelligence.
The difference that I see between the two is that, in the case of Apple there's no more data kept by the system than what you already have there today, but in the case of Microsoft it keeps everything that's on your screen and an analysis of everything that has ever been on your screen unless you disable Recall. So, if I don't explicitly save a file on my Mac or create a note in Notes or add a new synthesizer in MiRack, Apple Intelligence wouldn't have any idea I did anything at all. By contrast, a thing needs to merely appear on my screen for Recall to be aware of it. So I could open a Notepad window and start typing and that's now saved for eternity (or until I run out of Recall space) without needing to even save the note.
I think both systems sound neat and I'm excited to see where they go, but I am much more apprehensive with having an actual history of everything I've ever done on my PC than I am with simply having a supercharged search mechanism into what's already there. That's not to say that Apple's system is without issue. I still haven't processed how this might affect victims in abusive relationships for example, but in my use case and for people in my family I see Apple Intelligence being safer than Recall specifically because
- there isn't a centralized database that can be uploaded when we get a virus
- there isn't a database that contains all of our passwords and credit card numbers scraped from every entry field everywhere
- there isn't a stupid screenshot OCR tool running in the background, wasting power all the time
And, as an aside, on that last point from what I understand from the Apple State of the Platforms presentation, when you activate Siri/Apple Intelligence it'll analyze the screen using the actual UI components. So it knows there are text fields, images, and windows etc. It doesn't take a screenshot and OCR everything, but rather it uses the actual object tree so there's no lossy conversion nonsense where it tries to understand low contrast text somewhere -- it just gets the text from the presentation layer directly. When I heard Microsoft is probably just running everything through OCR, it hurt me a little. Apple's implementation isn't like some super fantastic implementation or anything -- it's just correct. Microsoft's implementation is just bad. I can only hope there was some huge hurdle preventing them from just grabbing the text that was already there because running screenshots through OCR just seems wrong. Especially when you're doing this every few seconds.
So for me personally and for my threat model, I'm fine with Apple Intelligence but I'll probably only keep Recall on systems that I use for like beta testing operating systems or whatnot. It won't be on my daily driver. Of course that all changes if we find out Apple feeds our data into ad targeting systems or Microsoft makes strides in blocking out sensitive data and truly securing their database. So we'll see how they develop over time.
-
Comment on Rumor: Insider claims Xbox handheld under development in ~games
snappyl Link ParentYou'd probably have to change your graphics settings every time you dock/undock, but this sort of set up does already exist with the Asus ROG Ally. They have a few eGPU you can plug into the ROG...You'd probably have to change your graphics settings every time you dock/undock, but this sort of set up does already exist with the Asus ROG Ally. They have a few eGPU you can plug into the ROG XG port or whatever it is and add a mobile RTX 4090 if you like.
Theoretically, you could also do something similar with upcoming handhelds running Intel internals that I suspect will expose a thunderbolt port.Downside though, at least for the Asus solution, is it is ludicrously expensive. I've made whole PCs to game on for the same cost as their eGPU.
-
Comment on Starfield - what are your thoughts? in ~games
snappyl (edited )Link ParentIn my experience on my beast of a rig, performance has been good and I have had no crashes yet. I'm only in about 8-ish hours. I'm seeing over 100fps most of the time on a 13900kf with a Radeon...In my experience on my beast of a rig, performance has been good and I have had no crashes yet. I'm only in about 8-ish hours. I'm seeing over 100fps most of the time on a 13900kf with a Radeon 7900xt on ultra 1440p. One thing that could potentially get a bit annoying are load times, though. I have a crazy disk array setup so my load times are around a second or two, but for comparison my XBox Series X takes 3-5 times as long and over a play session does get a little annoying if I'm doing a lot of zone hopping. That'll probably calm down though if I keep playing and learn where things are better.
Performance-wise on an XBox Series X, the framerate holds 30fps most of the time, but it'll have occasional dips. I haven't tried any gun battles on XBox yet -- I just did some shopping -- but I would expect they would be fine and hold steady at 30 the whole time. Graphics quality is obviously cut down, though. A hazard of the Zen 2 architecture and 16GB of total system memory I would suspect. Still, equally enjoyable as my PC I would say and definitely playable. Also I experienced no crashes here either.
For giggles I'll probably also try the game on a 13400f with an Intel Arc A770 16GB in the coming days. Should prove interesting.
Edit: It's unplayable on a ROG Ally on low 1080p. In the lodge I get a VERY inconsistent 10fps. Once leaving the lodge I get a crash to desktop.
Oh. My God. You just unlocked a new horror for me. Like I like using chat bots for trying to understand the current scientific understanding of their harms, and they also came in clutch this weekend finding cat feeders for my family and reducing my parent's cable bill. But this thought just makes me want to move to a cabin in the woods and live without electricity.