Sign in

Johann Beukes

@johannbeukes.com
142 followers 521 following 73 posts

👨‍💼Fractional CTO 📊 Data Scientist 🧑‍💻Startup life 😇 Angel investor 🥋 Martial Arts FL☀️R🌴DA living JohannBeukes.com

PostsRepliesMedia
Johann Beukes @johannbeukes.com · 28/04/2025
Reading while waiting for next flight, returning from South Africa (Stellenbosch) and it was amazing how this book from #PieterDuToit was recommended to me by both people who thought it was revealing a dark secret society and people who thought it was a badge of honor.
010
Reposted by Johann Beukes
Agentuity @agentuity.com · 05/02/2025
Generative AI has its challenges, but AI-augmented UI offers a promising near-term solution. Check out a prototype created with React and OpenAI from our own @rblalock 👇 www.agentuity.com/blog/ai-aug...
agentuity.com
AI Augmented UI: A Paradigm for Human-Computer Interaction - Agentuity
You hear a lot about generative UI but what about AI augmented UI?
021
Johann Beukes @johannbeukes.com · 03/02/2025
Not really. He doesn’t know what a “Braai Broodjtie” is. So…
020
Johann Beukes @johannbeukes.com · 03/02/2025
Last I heard from him, South Africa was part of the Sh!7 h0l3 countries… I guess this will be the next target. Renamed to … Dark Mexico. 🤷‍♂️
000
Johann Beukes @johannbeukes.com · 02/02/2025
That seems relatively low doesn’t it?
000
Johann Beukes @johannbeukes.com · 01/02/2025
Yes. Can’t go wrong with these lads.
010
Johann Beukes @johannbeukes.com · 01/02/2025
Am I one of the 2 monthly listeners? 🤷‍♂️ #AnalogSociety
010
Johann Beukes @johannbeukes.com · 26/01/2025
Well, that settles it. No brainer, you’re the cool uncle who let them go 😉
110
Johann Beukes @johannbeukes.com · 26/01/2025
I say let them go. Experience is a good teacher and/or could make for a great story. 14 seems old enough. *I’m not a parent
100
Reposted by Johann Beukes
Melanie Mitchell @melaniemitchell.bsky.social · 23/01/2025
Worth also pointing out that there are many "tests so easy no AI system can pass them". Moravec's paradox remains. E.g., arxiv.org/abs/2404.12390
arxiv.org
BLINK: Multimodal Large Language Models Can See but Not Perceive
We introduce Blink, a new benchmark for multimodal language models (LLMs) that focuses on core visual perception abilities not found in other evaluations. Most of the Blink tasks can be solved by huma...
711534
Johann Beukes @johannbeukes.com · 19/01/2025
Sunday in Florida. Ran into this good boy. So famous, people have tattoos of him.
000
Johann Beukes @johannbeukes.com · 10/01/2025
This week, I saw the power of AI assisted coding in practice. And it was glorious. - 3 teams - 3 projects *given to them* - 3 meet AND exceed results - only 24 hrs (instead of the normal 48hrs + 4hrs for demos) Give smart people all the tools they need!
000
Johann Beukes @johannbeukes.com · 10/01/2025
I understand that frustration, but to be fair to them, there are areas where their skills serve an important role. However, I very much would allocate the reason for success to: 1. The specific MBA program (school) 2. The person I have two partners in Venture with MBAs and they are phenomenal
000
Johann Beukes @johannbeukes.com · 10/01/2025
Had an awesome start of year kickoff with one of my clients this week, culminating in a 24hr hackathon between three teams. Seriously, I felt like a father, proud of every single one on the team. They exceeded expectations by a mile. The #whoop data supports the stress and effect on sleep
000
Reposted by Johann Beukes
Kirsten Lum @machsci.bsky.social · 28/12/2024
The #1 skill for delivering ML models at an eye watering pace is setting performance goals. The goal is not to build the best model. The goal is to build a model that beats the status quo, and get it in production as fast as you can. It works like this:
2373
Johann Beukes @johannbeukes.com · 29/12/2024
Btw… Shrinking is such a great show.
020
Johann Beukes @johannbeukes.com · 29/12/2024
Also Married with Children. I remember him as the neighbor.
140
Johann Beukes @johannbeukes.com · 29/12/2024
I had a great conversation tonight with a close friend and the topic of friendship came up. TLDR: A close fiend is not the one you call up when 💩 happened It’s the one you call to celebrate with you when something great happened, and there’s pure happiness coming from that friend with no jealousy
000
Johann Beukes @johannbeukes.com · 27/12/2024
I agree. Writing helps me organize and refine my thoughts for more valuable outcomes. I also prepare (read, research, etc.), which is why it works for me. However, some folks I know prefer to wing meetings without prep, they’d likely have a different perspective.
000
Johann Beukes @johannbeukes.com · 27/12/2024
100% agree. This past year, I’ve engaged and hired folks that: 1) I’ve worked with before this knew what they’ve done firsthand or 2) Had incredible experience with the specific technologies needed for a project. What I’ve NOT hired for was some formal education.
010
Johann Beukes @johannbeukes.com · 27/12/2024
Before reading this article I thought I’d have a strong argument against the statement, but not that sure now Yes, probability is subjective and an emergent property, but that doesn’t make it less useful. I’ve never thought of “true” vs “non-true” probability. I feel I’m missing something
121
Reposted by Johann Beukes
Sung Kim @sungkim.bsky.social · 27/12/2024
Apparently, you can run DeepSeek-V3 locally, provided that you have 8 M4 Pro 64GB Mac minis. ~5 tok/sec.
27911
Johann Beukes @johannbeukes.com · 24/12/2024
Ah yes. It’s seems even our local municipality got into some Festivus pole action 😉 A little too decorated, but they’ll learn the lessons over time that less is more, simplicity and rejection of all holiday excess is what we all seek… and of course the settling of grievances.
010
Johann Beukes @johannbeukes.com · 23/12/2024
Happy Festivus! (for the rest of us) “The Festivus pole symbolizes simplicity and humor, rejecting holiday excess with its plain, unadorned design. Its minimalist presence highlights the practical, parody-filled spirit of Festivus, offering both humor and reflection.”
110
Johann Beukes @johannbeukes.com · 20/12/2024
This seems like a good Friday before Christmas break research project. Thanks Tim. 😉
010
Johann Beukes @johannbeukes.com · 19/12/2024
So I love the fact it was able the response was able to identify the humor in the riddle by recognizing the lack of logical connection between its parts and highlighting how unrelated elements are intentionally juxtaposed for comedic effect. That’s some level of insight beyond next best character
100
Johann Beukes @johannbeukes.com · 19/12/2024
300 characters don’t allow full explorations, but the point I was attempting to make was that while we have kids in Forster homes waiting for adoption, that should be a serious consideration for folks who’s idea of kids is that the future depends on that. That’s all I was trying to say.
010
Johann Beukes @johannbeukes.com · 18/12/2024
Do you have any historical references perhaps we can lean on for some insight? Also, do you think there’s a path for “converting” some of those conflict seeking voters into Democracy seeking voters? We’ve all seen friends and family who voted for Obama or Bernie move to the other side and vote for 🍊
000
Johann Beukes @johannbeukes.com · 18/12/2024
How long is long enough to wait… Probably just skipping this update at this point.
000
Johann Beukes @johannbeukes.com · 18/12/2024
We waited too long to start talking about kids. We got married at 20, came to the USA with our suitcases three days after we got married, then worked as immigrants with no family to build a life in a new country. That’s why in our 30’s we had “the talk”. Maybe in our 20’s we would’ve had kids
010
Johann Beukes @johannbeukes.com · 18/12/2024
We share a lot of interest, but we also have very different interests which we appreciate and support each other in. I know it’s a 1 in “large number” chance to find someone who’s both your best friend and companion, but to me, that’s what a “life partner” is.
010
Johann Beukes @johannbeukes.com · 18/12/2024
Hot take: several of my friends with kids, have revealed if not fully admitted that some of the main reasons they had kids was because of 1) they’re legacy and/or 2) they don’t want to be old without having kids to look after them.
010
Johann Beukes @johannbeukes.com · 18/12/2024
I do enjoy not having to get up for soccer games on a Saturday morning, sleeping in and watching podcasts in bed until noon. However, I do agree there’s enough people, not because of the effect on the earth necessarily, but adoption should be a much higher priority for many considering children
110
Johann Beukes @johannbeukes.com · 18/12/2024
My wife and I probably fall into the first category. Started dating at 14, approaching 50 now. In our 30’s came to the conclusion we enjoy each other’s company and didn’t want to introduce anyone else into that mix. I guess that could be hedonistic, but I think it’s just being selfish with our time
110
Johann Beukes @johannbeukes.com · 15/12/2024
That’s the one I got on Amazon. Figured the 1st edition was from 2000, an update might be worth it.
000
Johann Beukes @johannbeukes.com · 15/12/2024
Here in Florida, we just need to have Fox News mention it’s approaching and there will be guns, Falcon rockets…etc, aimed at the sky. I mean, we shoot at Hurricanes 😊
110
Johann Beukes @johannbeukes.com · 15/12/2024
Ordered. Thanks for sharing Alex
100
Johann Beukes @johannbeukes.com · 15/12/2024
If you find value in learning it like being more productive or just even enjoying the learning experience, do it. Speaking for myself though, I’ve always just learned a new PL because it was required for achieving some goal, and I found my enjoyment in that goal, rather than learning the PL.
020
Johann Beukes @johannbeukes.com · 14/12/2024
Growing up in South Africa, the equivalent of this was Bully Beef”… mystery meat.
000
Johann Beukes @johannbeukes.com · 13/12/2024
New reading 📖📖📖📖📖 I try to keep a 2-1 ratio of reading DS/AI/ML and Brain/Neuro/Psy literature. #BrainScience #NeuroScience #Psychology @deanbuono.bsky.social
031
Johann Beukes @johannbeukes.com · 11/12/2024
I’ve always tried my best to make sure when talking to clients and coworkers that we appropriately talk about predictive analytics vs. machine learning vs. artificial intelligence. However, I’ve had not just one, but many pushbacks from marketing and sales wanting to call everything AI.
020
Reposted by Johann Beukes
Ron Itelman @ronitelman.bsky.social · 11/12/2024
I was informed I should share the YouTube livestream link, and not the zoom link I had previously shared, so here's the updated info: I will be sharing my research on innovation at the intersection of UX, AI, and psychometrics. Friday, Jan 17th 2025, 2-3PM MT www.youtube.com/live/kCmC4i-...
A promotional image for a tech talk about the intersection of UX, AI, and psychometrics with Ron Itelman with two parts. The top section shows event details: a headline 'Designing intelligent systems for trust' above a circular profile photo showing a man in a white shirt against a light background. Next to the photo is text 'You are invited! Please join my Active Inference Institute Gueststream' followed by the date and time 'Friday, January 17th, 2025, 2:00-3:00 PM MT'.
Below this is a comic strip with two stick figures in conversation. The first figure, with a smiling face, says 'YOU JUST TELL AN AI AGENT WHAT YOU WANT & IT'LL AUTOMATICALLY DO IT!' The second figure, with a deadpan expression, responds 'BRUH. YOU CAN'T EVEN EXPLAIN WHAT YOU WANT IN A STAND-UP.' The comic has a simple style with figures drawn in black lines against a light blue background, and dialogue in hand-drawn text within speech bubbles. Artist signature 'Ron Itelman' appears in bottom right. Ron has expertise in personalized learning and analytics for personalized knowledge and AI systems.
021
Johann Beukes @johannbeukes.com · 09/12/2024
😬 I need more than luck, I need some strong 💪 juju. 😊
010
Johann Beukes @johannbeukes.com · 09/12/2024
I found out today from one of my clients that they no longer have access to the training code for a specific model, trained using PyTorch and NTLK. So, they are asking me to come up with a strategy to "reverse engineer" the original model...oh, and yeah, I also don't have the original labeled data
110
Johann Beukes @johannbeukes.com · 09/12/2024
Ditto…
000
Johann Beukes @johannbeukes.com · 08/12/2024
Sunday read. 📖📖📖📖📖 Biocosm by James Gardner is a wild take on how life and consciousness could be central to the universe’s purpose. It’s definitely speculative and up for debate (with plenty of counter-theories out there), but it’s worth a read to stretch your perspective on the big picture
000
Johann Beukes @johannbeukes.com · 08/12/2024
Tree is up! 🎄🎄🎄🎄
000
Reposted by Johann Beukes
Sung Kim @sungkim.bsky.social · 07/12/2024
So... what is the best model if you want to generate a synthetic data? They looked at new generation, response generation, and quality enhancement. - GPT-4o excels at generating new generation - GPT-4o excels at response generation - Claude-3.5-Sonnet is better at quality enhancement
1152
Johann Beukes @johannbeukes.com · 07/12/2024
Any vids or write ups you’ve done on it?
110
Johann Beukes @johannbeukes.com · 04/12/2024
We can only hope. 😁
000