I decoded a viral explainer formula thecognitiveshift.com/public…
You can test it for yourself at the end, or just jump straight to the fun stuff (and read how it was done after) thecognitiveshift.com/traini…
Getting somewhere. Baby steps.
But tested outputs in FB groups and got good engagement on rawest form possible (text). Will be interesting to cook more with video and animations.
recently i have seen a few people fall into the trap of thinking they have an eval, try it in production, and give up on the project... all without training their judge 🤦♂️ .
figured I'd show people how (and build a tool for it *Grader*).
here's an example of building an eval for a funny comment bot you can run on a 27b local model in your basement... thecognitiveshift.com/public…
join bro and have a bit of fun by testing it here thecognitiveshift.com/traini…
if you want to learn how to make evals/benchmarks, make sure to signup to get notified when the course pre-order drops.
thecognitiveshift.com/traini…
msrp is $999 (but limited # of pre-orders will get it for $299)
I trained an AI judge on my sense of humour, then let an optimiser write comments to please it.
thecognitiveshift.com/public…
Nothing ground breaking. But since I'm building a tool (Grader) to build and test evals, I needed something quick to start with.
Now comes the fun (harder) eval... a remix on VendBench2.
Using my eval builder to make a funny reply agent.
Just a heads up it will be kinda unhinged.
Will be the intro example for the course I'm making on evals before moving to one I'm legit excited about.
thecognitiveshift.com/traini…
PS: course will come with access to a tool I'm building called Grader. Already finding it useful.
Sent the swarm out to gather data on the Pokemon market. Tons. This is tip of the iceberg. Also scanning Magic the gathering. Already learned some things I think Rudy is probably only half aware of if some of the questions he asks on lives are genuine and not market manipulation.
What a time to be a market participant. I still believe that even with all this insight, it’s a trap to over estimate ones edge. Or that it’s easy to loose sight of it.
It used to be “the” thing during the 1999 peak. Since then it’s one of the things. Adoption within younger groups are struggling and millennials and GenZ can only absorb so much. They’re printing +10billions of cards per set. New sets launching every 3 months and they’re expanding with the new plant live for 2028.
Look at the PSA number for mega evolution and it’s like in the tens of thousands per month vs a couple in previous eras.
DGMW. If bag holders double down positions as these print runs get bigger, even with demand the same, you can see the squeezes clearly. And if some escalation of conflicts creates black outs, games and cardboard will be it. You also have a baby boom happening within a millennial generation that is older, looking at mental health studies of kids these days who had too much access to phones… so there’s definitely many cases where even with higher print runs, the money keeps chasing.
But let’s be real. Look at what’s happening with the kids toy market overall. First those stuffed animals and objects (Jellycats). Then figgies (or however you spell them). Companies are rinse repeating the pokemon formula and its working. Too well.
Disguise a unboxing lottery into a child’s toy and you make more money selling to the speculative millennials than you do to the base use case. It’s quite smart how they do it. Sponsor a child’s YouTube channel to mention it. Small stock in brick and mortar and get some influencer to unbox and live react to a rare one and then buy a few rare ones back at crazy prices.
Kids see them in shops and react. Sending parents to tiktok to check. Who then find the unboxing and resale value and boom. Hooked. Suddenly they buy a few packs to unbox themselves and few more to store. And the produce goes from kids toy to speculative instrument in a flash. Kinda genius.
But can’t help but wonder when scalpers will need to learn the difference between real scarcity and the illusion of it.
Don’t know where I was going with this. But thanks for getting this far 🤣
DM me if you want some alpha on cardboard markets.
I see this and think it’s definitely a top 🤣 nitter.cf/lugiaconsulting/status…
This is the “we’re getting an ETF and institutional money is coming when BTC first hit 20k” moment.
Wait till people start burning cards on lives they lost so much money. Then consider DCA if you like what the adoption data shows in younger demographics.
graded pokémon just crossed $13.17b in total market cap.
sounds huge. look at what it’s sitting next to:
real estate ~$390t
bonds ~$140t
global stocks ~$120t
gold ~$30t
crypto ~$2.8t
art + collectibles ~$1.7t
now the thought experiment. not 1%. not 0.1%. just 0.01% of each rotating in:
~$39b from real estate
~$14b from bonds
~$12b from stocks
~$3b from gold
~$0.28b from crypto
~$0.17b from art
that’s ~$68b = 5x the entire graded pokémon market as it stands today. from one hundredth of one percent.
the capital hasn’t shown up yet. dyor
Ignore the contents of these, but just want to illustrate.
I have never seen anything pokemon related on X. Posted this thread and on next refresh I saw 5 in a row (if I ignore the ad in between).
This is what happens within the child toy category where companies roll the Pokemon playbook.
1. sponser YTer (educational channel aimed at kids)
2. sponsored TTer unboxing reaction (aimed at parents)
3. add supply in brick and mortar
What happens?
Kids see in store and react.
Parents search on Tiktok.
Algorithms boost mindshare.
Demand + supply shortage = price spike.
Price spike = WOM and headlines
= next time you make more of the stuff and still have buyers.
They have it down to a science.
And Pokemon was the spear. Wild.














