I had to rant about it. Sorry...

Please next time in a bathrobe

Thought about it 😅

The most annoying kind of posts are "it's so over for xyz" every time the new AI tool/model drops.

i have a strong feeling you are referring to @theo lowkey 😂 i still love @theo though

@theo Actually not referring to anyone specific. And theo‘s definitely not one of the „it’s over for xyz“ posters. He typically backs his statements up with facts. Though my first 12h with gpt 5 did indeed not match the experience he shared about it.

@theo No problem Max Also love you too… you taught me almost everything i know in my careeer Vuejs, Laravel, Typescript, Python, Go Student of your udemy channel and videos since 2017

@theo Wow, that's a long time! Thank you so much! :-)

They all want to be the next Elon, instead of actually making an impact in coding and tech as you do in React daily. That's why I respect you more than random influencers who can't even build a single app on their own.

Thank you so so much - this honestly means a lot to me! :)

Well deserved, you do make a difference in this industry.

🙌 Preach! Really refreshing to see this! I've seen dev-fluencers tweet > used GPT5 for weeks, it's not as good as sonnet 4 for coding Then literally 10 minutes later > It's an absolute game changer for coding! Anything for the hype engagement I guess🤷‍♂️ Never change Max! 💙

I agree, it doesn’t feel like a huge leap, especially compared to Claude 4. It also overthinks and gives the same output Claude would, just much slower.

I did the testing and came up with a similar conclusion. For existing code base Gemini 2.5 pro, Claude 4 >> Gpt 5 I'm just disappointed it is not that all the hype people are giving it to be.

Yeah, I did truly hope it would be that major step up. It Just isn't for me unfortunately.

"AI influencer" great term dude! 😂

At this point new models are only small improvements from previous models, these applies for all the models not only open AI

Absolutely

🔥

At first when they were released it was revolutionary now the new updates are marginal and incremental, I get the impression we have reached a plateau. Don't hold your breath AGI.

Thaaank you max thank you. I have been wanting to say the same thing after watching their demo, we need to see more than just the words here.

Man... exactly what I thought. It just exposed how many AI influences are grifting and hype hopping

The bigger influencers are paid, the smaller ones just copy. We need to ignore the hype and actually try extracting value from these models.

Thanks for the sincere review

Lately, this is where I am at with this as well. I LOVE AI. I think I am very, very AI forward. I also regularly consume technical AI content, but I am exhausted from seeing stuff where it's like "AI IS GOING TO TAKE OVER!" and it's just not. It's just very disappointing to see it every single time a new model version comes out, and my social feed turns into a sea of benchmarks. AI is great, don't get that twisted, but to constantly harp on it because that's where the views are drives me crazy right now. There is so much value and good content in the world of AI; we could be highlighting and showcasing more besides this. Fear and hype are what this creates, with nothing substantial to back it up. I use AI daily. I know it is a phenomenal productivity tool. So much can be accomplished with it, and you can leverage it to improve significantly, but the rest of the hype needs to be watered down a bit. The Kool-Aid you are drinking is just way too strong.

I think it's a good thing that people can still get hyped for things, but when it comes to LLMs it seems every new model that comes out is 'insane', 'life changing' and 'amazing' until the next new model is. This makes it hard for devs that can't test out every new model all the time, to keep informed and up to date on what really is going on in the LLM are(n)a.

Time to get back to JS frameworks :)

Guten Tag 👋

Hallo!

Yes, I have to filter out a lot of the noise, pay attention to feedback from people actually building and working with models, and most importantly come to my own conclusion through experimentation and experience.

> show don't tell Careful there! Last hyper hypester hyping up LLMs as a 60% productivity enhancer, 2 days to do work of 5, I asked to be shown evidence got worked-up into a frenzy ... was kinda sad, dude had to call in a buddy so the two of 'em could self-sooth one another.

Finally, some true advice. Just bragging that something is good doesn't always mean it's good 👍

Hey Max 👋

Hey!

Followed right away, appreciate the honesty, seems like everywhere you look people are overhyping it. I completely agree, it's a solid model, but absolutely nothing revolutionary there

Totally agree. I’m finding it really “nice” for brainstorming convos while I’m sitting on the couch dreaming about the next big thing. But for real work in Cursor, haven’t noticed anything above Claude 4 Sonnet

Thanks for saying it!

Don't be sorry, Max. You're entitled to your opinion even if it's not a popular one. The LLMs are just like the iPhones. When the first one came out, it's a huge breakthrough and revolutionary. But subsequent updates only bring incremental improvements.

Thank you!!! I tried GPT 5 and it was really good at analyzing my codebase and pointing things out, but tool calling was really not good. I have been back on Claude, I think for now it’s better, especially around tool calling.

You and me brother. We def have “this changes everything” ai-fatigue by now. Idk. Im gonna spend some real time with it maybe next week. The hype train has gotten annoying as hell.

well said man, the hype is really annoying

peoples are no longer buying this hype anyway. only those non-tech guys would fall into this trap.

I still can’t even use GPT 5 as a pro user. I’ll just stick with Claude for now. I haven’t used ChatGPT for coding since maybe 2024 except on some specific problems.

The hype isn’t hyping anymore. The dopamine hit isn’t the same as it was when GPT-4o came out. Now, it all feels the same.