396: Meta’s LLaMA, Baumol Effect, YouTube, Building AI Analysts, Doug Leone, Virtual Kidnapping, Visceral Fat, and Last of Us
"Do you want to be a cog or a lion? It’s hard to be both…"
Self-control: The secret to success as an investor.
Within you lurk an angel, a devil, a scholar and an idiot. If the angel and the scholar ever let down their guard, the devil and the idiot will wreak havoc that will take years of work to undo.
🚌🖍️👨🏻🏫🎓 The industrial-age model of education attempts to turn people into machines.
This is because industrial machines required operators to be present for the same amount of time every day, doing the same tasks over and over.
But human creativity is all about cycles, ebbs and flows, naps and sprints, finding the right mix of inputs and outputs, exploration and creation.
Do you want to be a cog or a lion? It’s hard to be both… ⚙️🦁
📞 🙏 👏👍 Back in Edition #389, I wrote this:
Some years ago, I decided to take the time whenever possible to tell people when I think they’re doing a good job. Many others may *think* it, but in my experience, few people *explicitly say it*. [...]
Making it specific moves it from polite social norms to something that the person will remember — maybe it’ll make their whole day. How would you feel if someone did this for you?
I want to keep promoting this idea and make it more common, so I’ll briefly share my latest experience with it.
We’re replacing our hot water heater. It’s a conversion from natural gas to electricity (there’s lots of cheap hydro where I live), and two technicians came to estimate the parts and labor.
I thought they did a great job, so when I called to make an appointment for the installation, I made sure to tell the person at their company’s office that I had feedback about them: “I think they’re doing a great job, they’re polite, answered all my questions, and went out of their way to address my concerns about certain things…”
The person on the phone seemed really surprised at first, and then very happy to hear positive feedback (for a change?). She said she’d relay the message to their supervisor and to them.
I hope it made their day better and gave them well-deserved pride in doing a good job.
🛀💭🧠 Sometimes I wonder if the estate of Albert Einstein ever thinks: ‘Could we please stop putting grandpa‘s name on random things to pretend that they’re smart please?’
🤧 I’ve been keeping score for months, my tally was getting to an all-time record of 8 or 9, but my streak of luck is over. I finally caught one of my kids’ colds.
I don’t know how I was able to dodge or fight off every single one of their viruses since last fall, but I hope my immune system can keep it up because it was great!
🏦 💰 Liberty Capital 💳 💴
🦙 Meta’s LLaMA is on the loose — AI model leaked on 4Chan 🤖🏴☠️
Facebook’s large language model, which is usually only available to approved researchers, government officials, or members of civil society, has now leaked online for anyone to download. The leaked language model was shared on 4chan, where a member uploaded a torrent file for Facebook’s tool, known as LLaMa (Large Language Model Meta AI), last week.
This marks the first time a major tech firm's proprietary AI model has leaked to the public.
This is the model with all the weights, not just a binary blob to do inference.
The package includes 4 sizes: LLaMA-7B, 13B, 30B, and 65B parameters.
Keep reading with a 7-day free trial
Subscribe to Liberty’s Highlights to keep reading this post and get 7 days of free access to the full post archives.