Empowering Veterans Through Code | Software Engineer & Educator | Ex-Microsoft | LinkedIn Learning & Frontend Masters Instructor | Founder @ Vets Who Code
The AI Peanut Gallery finally said the quiet part out loud. We are nowhere near AGI. Just yesterday, I had to step in and assist an AI agent that started with GPT-4.1, then switched to Claude 3.7, and still couldn’t figure out how to animate a basic emoji. It kept forcing a Tailwind-based solution when a simple CSS keyframes approach was all that was needed. It even left behind extra files it created but never used. No cleanup. No understanding of scope. I had to step in and finish the job. This isn’t intelligence. It’s a really great intern. Fast, helpful, capable in the right hands — but still needs supervision, clear direction, and someone to clean up afterward. I’ve said it before: AI is a force multiplier. It scales the ability of the human driving it. When I’m working in my strengths like front-end or Python, I’m moving through a quarter’s worth of tickets in just a few weeks. But when I use it in a space where I’m weak, like Kotlin, all it does is help me write bad code faster. The real headline here is that AI won’t replace engineers. It will reflect them. And companies still need people who know their vertical, their stack, and their edge cases. https://lnkd.in/eSNkHqYQ
Chris Hood
2d
I agree, we are nowhere near close to AGI, despite the hype. Here's a perspective I wrote this morning on AGI, and how we should start thinking about modeling it. https://www.linkedin.com/pulse/desire-agi-moving-beyond-commercial-hype-measurable-chris-hood-kr3ac/
Supply a ton of information, precise and descriptive tasks for something small and it can really be helpful. Outside of this, can get sketchy. Don't vibe code.
GenAI won't replace writers, producers, artists, engineers, or analysts it will help them level up or it will diminish their capability due to over-reliance. It's a great research assistant, secretary, and can be adept with summaries with the right prompting. As long as we're designing systems with a motivated and attentive human in the loop the possibilities are endless. But without a human with both qualities, it's led to costly and embarrassing errors.
Sami Gardner
1w
CEOs really want to cut payroll costs so they will ignore anything that says they can't replace employees with an LLM/Gen AI. There are many people who want to make money on the AI-rush who will hype up LLMs. It's the next tech bubble. I say this as someone who uses OpenAI everyday.
Mike Logan
1w
Try getting AI to illustrate Genesis 21 or Genesis 22. Very eye opening on how the models are trained to be politically correct, which is ultimately artificial intentionality…the art of make believe. I’d love to hear about models that just do what we ask without rationalizing who will be butthurt if it illustrates truth.
Founder of Ask Sage, Bringing Generative AI to Gov | Former U.S. Air Force and Space Force Chief Software Officer (CSO) | Pilot
1wThis will age well.