Roger Gentry
banner
r-o-g-e-r.com
Roger Gentry
@r-o-g-e-r.com
🦾 Crafting tech to make humans more superhuman. 👨‍💻 Cyborg architect by day, 🐾👶 servant to pets & tiny humans 24/7. 💫 Chaos → Innovation
It’s the same grind a jr employee would have to experience and iterate on to become effective at their role.
October 31, 2025 at 3:58 PM
You look at the company books and labor costs vs completed outcomes?

Every industry is different but you have to establish a metric of success based on the current baseline then compare what changes.

Labor/time to complete task/time to redo tasks.
October 31, 2025 at 3:46 PM
Cost of licensing X number of users = monthly investment costs.

Then look at the increase in revenue and cost savings before / after implementation. Does this exceed the total investment into the license cost ?

Problem is most businesses have a hard time giving the base KPI to measure from.
October 31, 2025 at 3:37 PM
The promise: “your company will be faster and smarter!”
The reality: multiple thousands per month spent on a license for a Ai software that only like 3 people of 300 actually produces any ROI from the tool.
October 31, 2025 at 3:28 PM
I have a lot of questions…
October 29, 2025 at 6:13 PM
I feel like a Pringle chip is a great representation of LLMs.
October 28, 2025 at 3:03 PM
IMO if you’re crawling for commercial use you should be paying to organization for their data.
October 27, 2025 at 2:28 PM
I’m building these solutions and see the results. For some resources research.nvidia.com/labs/lpr/slm...

www.ibm.com/think/insigh...
Small Language Models are the Future of Agentic AI
Project website for the paper 'Small Language Models are the Future of Agentic AI'
research.nvidia.com
October 19, 2025 at 2:11 PM
“Trust us we definitely didn’t hide a backdoor”
October 18, 2025 at 3:51 PM
We are rapidly approaching the point where as an individual most of your ai usage will happen on device and not in a massive datacenter.

Most of the general practical applications with this generation of LLMs can be done with models that can run on consumer hardware with less power than an Xbox
October 18, 2025 at 3:50 PM
Especially in the shower so you know if someone is hiding behind the shower curtain waiting for you.
October 12, 2025 at 3:33 PM
In my mind I can only read that as “deep farting” for some reason. Yeah that needed to be shared.
October 12, 2025 at 3:31 PM
There’s a convenience factor it’s almost worth it.
September 11, 2025 at 1:59 AM
That context length will really help it compete.
September 5, 2025 at 7:36 AM
This is just the evolution of the script kiddy.
September 5, 2025 at 12:42 AM
Have you seen Google’s SynthID: ai.google.dev/responsible/...
SynthID: Tools for watermarking and detecting LLM-generated Text  |  Responsible Generative AI Toolkit  |  Google AI for Developers
ai.google.dev
September 1, 2025 at 12:40 AM
In America 🇺🇸 you give consent in the technology acceptance form during onboarding.
August 31, 2025 at 10:07 PM
Feels a little to real reading this while waking into the garage past the cabinet of expired pant and chest freezer to do a load of laundry….
August 31, 2025 at 4:31 PM
Had to make a poster.
August 23, 2025 at 4:46 PM
Why be obscure?
August 18, 2025 at 5:33 PM
Just one more layer of defense? Common what’s it gonna cost a few gb of memory and another 40% of your cpu? You can’t not afford this security! We stop tens of dozens of issues yearly.!!
August 17, 2025 at 3:58 PM