Vibecoding: a beginner's guide

There was a brief moment when I was balls deep in some projects (for recruiters) & ran into Claude & ChatGPT limits/fokol tokens & was forced to jump onto Grok.

Claude was Sonnet 4.6.

I will still maintain out of my main threesome (Claude, ChatGPT, Grok) for technical aspects/requests Claude is unmatched.

For conversational aspects ChatGPT & for Twitter/x/xHamster referencing & deep diving it is Grok.

And yeah AI sycophancy is a real issue esp with ChatGPT, been burned a few times with job potentials :(
 
FiestaST said:
There was a brief moment when I was balls deep in some projects (for recruiters) & ran into Claude & ChatGPT limits/fokol tokens & was forced to jump onto Grok.

Claude was Sonnet 4.6.

I will still maintain out of my main threesome (Claude, ChatGPT, Grok) for technical aspects/requests Claude is unmatched.

For conversational aspects ChatGPT & for Twitter/x/xHamster referencing & deep diving it is Grok.

And yeah AI sycophancy is a real issue esp with ChatGPT, been burned a few times with job potentials :(
Click to expand...
?
 
FiestaST said:
Comprehension fail? My post was fairly clear.
Click to expand...
Nee. Was kak en kom voor soos dronk kots.

Clean up single pass

"For a while, I was juggling several projects for recruiters and hit the usage limits and token caps on both Claude and ChatGPT. That forced me to start using Grok as well.

At the time, Claude was running on Sonnet 4.6.

Having now spent considerable time with all three — Claude, ChatGPT, and Grok — I can say that each has a clear strength. Claude is unmatched when it comes to technical tasks. ChatGPT handles conversational interactions best. And Grok excels at referencing and deep-diving into Twitter/X content.

One broader concern worth noting: AI sycophancy is a real problem, particularly with ChatGPT. It has a tendency to tell you what you want to hear rather than what you need to hear, and that has cost me a few job opportunities."
Click to expand...

Edit
So 9pm updates are powered by grok :unsure: 😂
 
Last edited:
@RedViking

you may have this in settings -> usage

Anthropic Is Giving Away Free Claude Credits — But Only Until April 17​

If you use Claude Pro, Max, or any plan at all — stop what you're doing and go claim your free credits right now. Anthropic is handing out a one-time credit equal to your plan price, and the deadline is April 17, 2026. Miss it and the credits vanish forever.

I got $100 free. You might get $20, $100, or even $200 depending on your plan. Here's everything you need to know.
Click to expand...

www.ayyaztech.com

Claude Just Gave Me $100 Free — Claim Yours Before April 17, 2026

Anthropic is giving free Claude credits equal to your plan price — $20, $100, or $200. Claim before April 17, 2026 or lose them forever. Plus new usage bundles with up to 30% off.
www.ayyaztech.com

not sure how it works but I just claimed mine and the usage bar has not moved / resetted

edit
I assume its this kak

1776330997574.png
 
Herr der Verboten said:
@RedViking

you may have this in settings -> usage



www.ayyaztech.com

Claude Just Gave Me $100 Free — Claim Yours Before April 17, 2026

Anthropic is giving free Claude credits equal to your plan price — $20, $100, or $200. Claim before April 17, 2026 or lose them forever. Plus new usage bundles with up to 30% off.
www.ayyaztech.com

not sure how it works but I just claimed mine and the usage bar has not moved / resetted

edit
I assume its this kak

View attachment 1901319
Click to expand...
I cancelled the service. Ill check it out. I tried to used it today to check a Notion Page and my credits are already finished. Its useless to me. I understand now why they are taking all the ram and storage in the world. This thing is gonna crash big time.
 
New paper by Microsoft Research:

Our large-scale experiment with 19 LLMs reveals that current models degrade documents during delegation: even frontier models (Gemini 3.1 Pro, Claude 4.6 Opus, GPT 5.4) corrupt an average of 25% of document content by the end of long workflows, with other models failing more severely. Additional experiments reveal that agentic tool use does not improve performance on DELEGATE-52, and that degradation severity is exacerbated by document size, length of interaction, or presence of distractor files. Our analysis shows that current LLMs are unreliable delegates: they introduce sparse but severe errors that silently corrupt documents, compounding over long interaction.
 
Last edited:
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X