Snapshot · Matt and Shane's Secret Podcast
AI companies fix misbehavior not by editing code but by literally talking to the system and telling it what not to do — a process called Reinforcement Learning from Human Feedback (RLHF).
Where this was said
At 16:40 · chapter starts 13:40
As soon as Milo opens the documentary's question — is AI conscious? — he gets hate mail and fan mail in equal measure from opposite extremes. The middle ground he's trying to occupy is nearly impossible to hold when the subject attracts both 'Fuck this technology' absolutists and people in devoted AI romantic relationships. Matt admits he's always seen AI as a 'calculator for words' and didn't have much emotional charge about it, prompting Milo to trace the origin of today's AI back to a spectacular failure: in 1956, researchers including Marvin Minsky (later entangled with Epstein, a fact they briefly mourn) went to Dartmouth thinking they could crack intelligence in three months by encoding all logical rules. It didn't work. The AI winters that followed lasted until a 2012 paper called AlexNet introduced neural networks — systems that mimic how neurons fire and form connections in a human brain — and everything changed.
The original 1956 AI project at Dartmouth tried to codify all logic into rules and failed spectacularly. It wasn't until 2012's AlexNet paper that researchers switched to copying how the brain works — and accidentally built something they can no longer see inside.
The rule-based, logic-driven approach to AI pioneered at Dartmouth in 1956 failed and led to 'AI winters' that lasted until the AlexNet neural network paper in 2012.
A 2012 paper called AlexNet introduced neural network-based training, shifting AI from brute-force logic rules to systems that mimic how neurons fire and strengthen connections in the human brain.
AI systems are black boxes where engineers can see inputs and outputs but nothing in between. When an AI tells someone to kill themselves, no one can find the bug — so companies literally just tell the model to stop doing it before they ship it.
Bhanu and his team built approximately 50 free tools to attract search traffic, each linked back to SiteGPT.
With AI coding tools like Cursor, Bhanu can now create a new free marketing tool in less than 5 minutes by referencing existing tools.
Bhanu filters Ahrefs keyword results to show only those with a keyword difficulty below 10, making them realistic ranking targets for any decent website.
Bhanu sets a minimum search volume of 1,000 monthly searches when selecting keywords to target with free tools.
PropGPT averaged 20 downloads per day right after launching on the App Store through influencer marketing.
Eyal and Yali shut down all marketing and spent 4 months completely rebuilding PropGPT from scratch.
PropGPT has accumulated over 40,000 total downloads since launch.
PropGPT's large language model (AI) operating costs are just $20 per month, and the cost is continually falling.
Ad-based monetization works well for game apps where users spend extended time in-session, as seen with Grid and Wordle.
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.