Let me get this straight... we're on the verge of both:
1. Scientists discovering a cure for the disease of aging, so that we can live forever.
2. Engineers building AGI that they believe could kill us.
Isn't it ironic, don't you think?
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
36
I'm usually the guy that says "They overpaid!" or "Not worth it"
In this case: "@Cmakar8 is worth every penny" and in hindsight, may look like a steal. @Avalanche smart to lock this guy up. @NHL
sports.yahoo.com/nhl/breakin…
28
Interesting
I am going to start to share the leaderboard at VulcanBench more regularly.
VulcanBench is the only benchmark I know that benchmarks across effort levels, on an eval suite of 100% real engineering tasks.
Getting ready to sunset eval suite 3 and eval suite 4 is almost ready to go.
Here's the current leaderboard for Eval Suite 3, Grok 4.5 High is the highest scoring model still with Fable 5 Low, yes Low, right behind it.
What I've been able to uncover with VulcanBench is that way too many people are running models at Max effort, thinking that buys them more accuracy, when really it just costs more and uses more tokens so takes more time.
If you're still in the mode of, tell my agent to do something then go get coffee, you're probably still living in the past. You can move faster, with higher accuracy, the key is not thinking you need Max effort all the time.
Dad'ing, Golfing, and Common-sense thoughts retweeted
We talk golf, golf trips, and we laugh with @KylePorterNS. Listen to the show at: youtu.be/GAIndB9UIwA
The squirrel was still alive at this moment. Would you save the squirrel or let nature take its course?
22
New PR! I had a rough winter, learning. To deal with wet, soggy conditions. I'm swinging less (work) but I've found a good swing thought and mental game is sharper. My bad scores are all on *first-time" courses.
24
On a roll. We're back to the single digits after a rough winter and spring season. Driver is performing with more relaxed tempo. Irons holding up and putter is not holding the train back. How low can we go?
17
This video was produced from *1* single photo! 🤯🤯🤯
Company training videos are terrible. The On-boarding process at most companies is lackluster, to say the least. These processes need a complete revamp. This is an easy AI use-case that EVERY COMPANY should adopt. It's safe. All you need is 1 good photo and the product materials or context you want to cover.
Here is how the video below was created.
The main effort for this was producing the multitude of reference images from the single base image. But those are re-usable assets that can be used for all future videos.
Creating the actual video? ...A few prompts and waiting for generation.
The entire process took about half a day, but much of that time is waiting. In between waiting, I cooked dinner, ate dinner, cleaned up dog poop, watched a hockey game, and researched using Claude + Lemlist (which is pretty awesome, btw) for a client.
I did go a little extra mile with voice generation because I wanted it to sound more like me. So, I trained my voice model. And then the lip syncing adds generation time.
But now with much of the base work done, I will look to hook this up to an agent harness and create these programatically. I can imagine pumping videos out in minutes (minus generation time).
Want to create a video for your company training?
Or want to learn how this is done?
DM me.
[Aside: When I say these are wild times, these are wild times! What a time to be alive!!!]
64














