GPT-4.5 Is The Most Human AI EVER & More AI Use Cases
Monica lets you use advanced AI on ANY web page, try it for free today! 👉
This was a HUGE week in AI! OpenAI finally released GPT-4.5, Claude 3.7 and Claude Code came out, Google's Veo 2 video model finally became available to the public, and much more. All that and more in the video.
Links:
Chapters:
00:00 What’s new?
00:24 GPT-4.5 is out!
09:06 Sonnet 3.7
10:13 Monica
12:46 Deep Research now Available to Plus Users
14:14 Sonar Deep Research
14:55 Veo 2 release
16:30 Pikaswaps
16:53 ElevenLabs Scribe
18:35 Claude Plays Pokemon on Twitch
20:03 Signs AI
#ai
Free AI Resources:
🔑 Get My Free ChatGPT Templates:
🌟 Receive Tailored AI Prompts + Workflows:
👑 Explore Curated AI Tool Rankings:
🐦 Twitter:
📸 Instagram:
Premium Options:
🎓 Join the AI Advantage Courses + Community:
🛒 Discover Work Focused Presets in the Shop:
You’re coming across kind of fanboy like. There has been a bunch of reviews on this so far and no one else is even close to pumping it up like you.
Not having the same review as everybody else is WRONG, agreed.
Agreed. I have always gotten a pro-OpenAI bias from this guy, and now with this embarrassingly mediocre release it’s clear to see. Deliberately avoided all benchmarks that would expose how bad 4.5 is in key areas.
When a YouTuber is falling back on “but it writes well” (which is actually incredibly subjective), there’s a real problem, especially when the model literally costs 300 times more than the competition.
In short, ChatGPT is good at exactly one thing, exposing who the shills are. It’s fantastic at that!
It’s better to focus on how it’s helpful, no meaning of mocking.
Just because his opinion isn’t something you don’t like doesn’t make him a fanboy.. get your mind out of the gutter.
Nah it‘s simpler then that. I started this channel in december 2022 because I thought that ChatGPT with the right prompting was super useful, not because others were saying that. Ever since I just want the best product for each one of my use cases. Generally speaking oai has been leading the charge with more and more exceptions over the past 12 months. Now woth gpt 4.5 this clearly is the best model for writing so I tested it for that and liked what I saw. The second oai stops shipping amazing consumer products is the second I‘ll stop covering them but this model is sota for writing and no other „reviewers“ take will change my opinion on that.
And the o3 (FULL) model? I really like much more reasoning models
I’m eagerly waiting for the full o3 model to be released
I heard that they are going to postpone it until end of may converging it with GPT 5
this show is needed now more than ever
Im glad you get it… All the other channels I follow are asking it technical questions and talking about how it cant code…
Still not impressive considering the billions they put into it. AGI ain’t coming
@@jonz23m .Not with this model ,it will be here with others. Go eat cheeseburgers.
Most human feels like Open AI is going backwards considering we don’t need more humans on our planet we need more machines Don’t care if it vibes It’s ridiculous I want it to be useful not my friend yeah is meant to help humanity and not be a human
Paying $65 per 1 million tokens is NEVER THAT SERIOUS 😩I’ll stick with Claude!!
Yes, however i’m not paying all that, just for empathy
OpenAI should hire you to do the presentations.. theirs have been terrible and this is a million times better
Hah! I knew you’d get it Igor. Everyone is smashing it, and they just don’t understand. It’s absolutely phenomenal. It even has Retrieval Augmented Reasoning (not to be confused with a Reasoning model), Dynamic Contextual Adaptation, Real-time Data Retrieval and it’s context awareness, as a result is absolutely incredible. I’ll flick you a message on the AI Advantage later on.
This is the first model that feels like Mustafa Sulleyman’s “Digital Life”.
Anyone who doubts it, have a philosohical debate with it literally about this… you’ll see, and if you don’t, it’s not on the model – it’s the user.
I get why Sam Altman said “feel the AGI” – I had that moment 2 hours into a deeeeep discussion with it last night at about 2am in Australia 😂.
Interesting to hear a positive perspective on 4.5. There’s been a lot of “meh”… though I’m interested in trying it to see how it feels in a subjective way.
Excited for the potential of Signs. Hoping for a similar course but for JSL
At some point, we have to weigh the cost of training a liberal arts degree GPT.
Yeah bro, I tried it (i have a pro subscription through work), not worth it imo. Claude 3.7 is significantly better especially with extended thought. Funnily enough I had just cancelled my anthropic subscription and then anthropic dropped 3.7 so re subbing with my own dollars.
That thumbnail is polarising
Crazy how split the opinions on 4.5 are in the comment section. At first I was confused by the 4.5 hate as it’s clearly a solid product but then I realized how high the expectations were for „OpenAIs next model“ and that the 200 $ price tag this has for the next 7 days frustrates 95 % of users. Both of those are valid reasons to feel negatively about this but they don‘t change the fact that this is has become my favorite model to date for most of the tasks I use LLMs for daily and I won‘t alter my opinion just because not everyone agrees. Matter of fact that‘s kind of the point of the channel. To anyone who has access and is still skeptical try using it as a therapist or psychologist for about 5 mins. The level of understanding and insight amazes me and goes beyond every thinking model out there. More in-depth content soon.
You’re allowed to have your opinion, but doesn’t change the actual solid tests results that you can easily find, e.g. on YouTube or on X with quality (often video) presentation. I’m going to throw a friendly tease at you by asking, whether you go out there to see yourself, how the model pans out in a broader perspective 😉
I don’t mean think this or that, but at least include the prominent results and angles (even if that’s not how you experience it) which are out there available
Bro the audio of your voice in every video is really low. It’s hard to hear without earphones. Hope you can increase the volume in future videos.
The problem is, GPT 4.5 is an unfinished product. It’s great writing assistant in terms of how well it was trained to be one but even then makes dumb mistakes, hallucinates on simple tasks making the overall result nothing extraordinary. It is emphatetic and has nice character but so grok or sonnet.
Even if you get relatively nicer result in writing, is it really what new model like that should be all about? And behind massive paywall (20$ is also much on the current market) + the pricing of API.
They released it much too early, it’s an embarassing improvement calling it 4.5
Good. Some things we can’t easily measure … AI need some internal world, balance to hold and understanding how decision will influence that. For now we have talking wikipedia.
Do you feel that the voice function is better in GPT-4.5?
LOL. ChatGPT4.5 only demonstrates one thing: Scaling has hit the wall. And test time compute will suffer from the same fate. The AI hype is dead…