I can totally see why Anthropic increased the price of Haiku-3.5 over 3.0. If it truly has performance of Opus of even better, then it would be a no brainer to move from £15/$75 to $0.25/$1.25 per million tokens. At the emd of the day, Haiku 3 at $0.25/$1.25 is still available if price matters. As soon as Opus 3.5 hits, Haiku 3.5 will almost certainly revert to previous pricing.
Big fan of all your work! Thank you for these insights and your research as always. Can we also talk about the environmental impact of using chat bots, and how to use LLMs responsibly and in consideration of all the costs, including those that are externalized / hidden? Utilizing LLMs as a glorified/slower Google incurs major environmental costs, which I think deserve mention - especially as more info comes to light on the water resources required to keep these data centers afloat. My apologies if you may have already discussed this on your channel and I missed it!
I work for a company focused on water management. As we know, data centers use a lot of water. Recently, we were approached by one of the huge players in the AI industry. It was not to use our technology and services in their data centers. Surprisingly, they wanted to offset their impact by providing water management to schools, libraries, etc (at their expense). This is very encouraging and is an example of how their impact can be indirectly lessened.
Man I just clicked coz of the thumbnail...makes me think of an ideal life where I'd wake up every morning in shape, not tired, well shaved well dressed well hair-stuffed with an ideal desk with an delicious cup of coffee stuff with the ideal amount of sugar in the coffee (yes i use sugar in the coffee)...ah that would be so...ideal...Thanks for this amazing thumbnail, and if it's ai made then thanks to the AI..I just wanna live that dream of ideal mornings in the sun and coffee and stuffs..freaking yeah...
The real prize in AIs will be AGI. LLMs can be very useful but they won't lead to AGI. They will probably form a part of it but not the "general intelligence" part. That will require an architecture that can create predictive models. The only company I know that's working on this at the moment is Google. Early days yet but they're seeing some promising results.
Had a very interesting experience recently with 3.5 Sonnet. It was not following my instructions and I got irritated with it after about 3 events and I asked it why was it not following my instructions. It apologized and then continued it's task but with more bad behaver. After calling calling it out a second time for not obeying my instructions (I referred it back to the clear instructions in the previous prompt), it became very pissy, (if that's a real word). It was like an angry girlfriend mouthing off. Very strange! Anyone else experienced this?
I’m in the US and I don’t see a setting to allow Claude to read visual information like charts/ graphs. So far it can only read/understand the text portion.
Igor was denkst du wird das AI Thema auf die ppc Welt haben? Hast du dazu eine Meinung? Ich arbeite mit Google ads all day. Aber persönlich suchen mache ich auch mehr über die AI Plattform also über die Google-Suche. Ich würde deine Meinung dazu echt interessieren.
I noticed ChatGPT seems to have dipped in the last week or so, whereas it was impressing me with the rapid progress before that. This is ironically what led me to try Claude.
I appreciate you, but I don’t understand why you’re also following RUclipsr trend of cutting out every quiet moment in the speech. It’s really unbearable-how do people even have the patience for this?! I thought this trend would end at some point, but it’s only getting worse. We need to breathe, to actually listen and watch, and have a bit of human quiet…We’re not a machine.
Hey The AI , awesome video! Your unique style really stands out. I'm a video editing specialist focused on RUclips growth, and I can help enhance your content with high-quality edits and engaging thumbnails. If you're ever open to exploring new ideas or strategies that can boost your channel further, I'd love to connect and see how we can collaborate. Let me know your thoughts
I can totally see why Anthropic increased the price of Haiku-3.5 over 3.0. If it truly has performance of Opus of even better, then it would be a no brainer to move from £15/$75 to $0.25/$1.25 per million tokens. At the emd of the day, Haiku 3 at $0.25/$1.25 is still available if price matters. As soon as Opus 3.5 hits, Haiku 3.5 will almost certainly revert to previous pricing.
But its not, its in level of 70b llama 3.1 where it but better but worse then llama 405b
Big fan of all your work! Thank you for these insights and your research as always. Can we also talk about the environmental impact of using chat bots, and how to use LLMs responsibly and in consideration of all the costs, including those that are externalized / hidden? Utilizing LLMs as a glorified/slower Google incurs major environmental costs, which I think deserve mention - especially as more info comes to light on the water resources required to keep these data centers afloat.
My apologies if you may have already discussed this on your channel and I missed it!
I work for a company focused on water management. As we know, data centers use a lot of water. Recently, we were approached by one of the huge players in the AI industry. It was not to use our technology and services in their data centers. Surprisingly, they wanted to offset their impact by providing water management to schools, libraries, etc (at their expense). This is very encouraging and is an example of how their impact can be indirectly lessened.
Claude coming on top ❤🔥
You do an exceptional job in communicating and creating perfectly balanced videos and content. You’re a breath of fresh air.
much appreciated 🤝
Man I just clicked coz of the thumbnail...makes me think of an ideal life where I'd wake up every morning in shape, not tired, well shaved well dressed well hair-stuffed with an ideal desk with an delicious cup of coffee stuff with the ideal amount of sugar in the coffee (yes i use sugar in the coffee)...ah that would be so...ideal...Thanks for this amazing thumbnail, and if it's ai made then thanks to the AI..I just wanna live that dream of ideal mornings in the sun and coffee and stuffs..freaking yeah...
It's even more surprising to share my amazement at your ability to access the exchange rate and the weather.
The real prize in AIs will be AGI.
LLMs can be very useful but they won't lead to AGI. They will probably form a part of it but not the "general intelligence" part. That will require an architecture that can create predictive models. The only company I know that's working on this at the moment is Google. Early days yet but they're seeing some promising results.
Uhh the idea of ai generated games actually makes my decart
Great update. Thank you!
Woah, seems I'm late to the party! Guacmadocious, Igor.
phew I hoped to catch this comment. now I can sleep in peace
@@aiadvantage My notifications failed me... (you're the only channel I have them on for) have a good night :)
AI Sheldon cam see all types of images and pdfs located in any drive storages location.... Anthropic is months behind ;)
Had a very interesting experience recently with 3.5 Sonnet. It was not following my instructions and I got irritated with it after about 3 events and I asked it why was it not following my instructions. It apologized and then continued it's task but with more bad behaver. After calling calling it out a second time for not obeying my instructions (I referred it back to the clear instructions in the previous prompt), it became very pissy, (if that's a real word). It was like an angry girlfriend mouthing off. Very strange! Anyone else experienced this?
I’m in the US and I don’t see a setting to allow Claude to read visual information like charts/ graphs. So far it can only read/understand the text portion.
Sorry, on the Claude app I could not find this feature, but on the desktop version I see where you can change this function. Thank you!
I heard this was pretty guacmadocious, checks out...
can confirm haha
GROK does not censor anything, which makes it the best. The bias in others is horrendous
strong argument! but prompt adherence and world knowledge do matter
Is it possible to see/experience this digital twin..? Online/VR?
Igor was denkst du wird das AI Thema auf die ppc Welt haben? Hast du dazu eine Meinung? Ich arbeite mit Google ads all day. Aber persönlich suchen mache ich auch mehr über die AI Plattform also über die Google-Suche. Ich würde deine Meinung dazu echt interessieren.
is it just me or has chat gpt suddenly changed completely? he always answers the same and reacts much worse than before.
I had that too..
I noticed ChatGPT seems to have dipped in the last week or so, whereas it was impressing me with the rapid progress before that. This is ironically what led me to try Claude.
For university students you suggest Claude or chat gpt pro?
Still ChatGPT for the tooling unless you are studying computer science then Sonnet for the superb code generation
@3:22 Speech-to-text
Where in Brazil are you at mate?
This was shot in Guriu, Ceara :)
@aiadvantage I live nearby and work with AI. Small world
The 11 Labs X voice thing is a useless gimmick.
Agree. Was fun to try though
Erster
I appreciate you, but I don’t understand why you’re also following RUclipsr trend of cutting out every quiet moment in the speech. It’s really unbearable-how do people even have the patience for this?! I thought this trend would end at some point, but it’s only getting worse. We need to breathe, to actually listen and watch, and have a bit of human quiet…We’re not a machine.
Hey The AI , awesome video! Your unique style really stands out. I'm a video editing specialist focused on RUclips growth, and I can help enhance your content with high-quality edits and engaging thumbnails. If you're ever open to exploring new ideas or strategies that can boost your channel further, I'd love to connect and see how we can collaborate. Let me know your thoughts