Thinking about it

Category: AI

  • Asking GPT4 about Tolstoy

    I’ve been thinking about this conception of LLMs as a lossy compression of the internet. Ted Chiang talks about this here. Karpathy also mentions it in his latest video. What does this loss look like? It doesn’t always seem like the “loss” part matters. Maybe its even good. Maybe sometimes what you want is some…

  • New Karpathy Video

    I am a big fan of Karpathy’s style of explaining things so I was very happy to see this video drop. Honestly, Its not my favorite of his but it covers a lot of good topics. I enjoyed the second half more than the first but its still a good starter video for people who…

  • Latency, UX and LLMs

    Latency is well understood to be detrimental to the user experience. No one likes to wait. But there are multiple ways this affects the UX In some cases the latency of interaction can change our relationship to the entire process. We see this a lot in software development where the turnaround time from making a…

  • GenAI and Human Creativity

    A friend shared this pretty impassioned defense of the idea that ChatGPT is bad for humanity because lowers the value of creativity which is important to the “Very Soul of the World” I am not going to try to pick apart the argument. It is a reasonable one. I do want to try to think…

  • AI Safety

    I am all for making AI better and more aligned with things which are good for people. I don’t really like the term AI Safety though because it seems to encompass a lot of different things which need their own proper attention. Stuff like: These are all important but they range in practicality from Already…

  • OpenAI Playgrounds

    I knew about these but I hadn’t really been using them much until someone really showed me how cool they are for “playing around” with GPT. Its obvious when you say like that but I am often not very smart. When doing experiments with the chat interface I often want to try out different paths…

  • GPT4 Helps me Code

    Spent some time yesterday giving some of my crappy code to GPT4 and asking it to make it better. I gave it this: and asked “Can you simplify this code” The response was pretty good. Then I asked “Can you add type annotations?” And got back: It also added this note: Note: The trace is…

  • LLMs and Pocket Calculators

    I am old enough to remember when calculators became a thing. There was a weird feeling for a while that this little machine could do math so much better than people could. There was a lot of talk at the time about how smart will computers get. Will they be able to win at chess…

  • What can LLMs do?

    I wrote this up a while ago for a friend but I decided to update it and post it here. LLMs are capable of doing a lot of interesting and surprising things. They also sometimes fail in surprising ways. Most things I read online about LLMs are either people trying to say either: “Well they…

  • Image Rotation and Image Understanding

    I did a bike tour across the Olympic Peninsula this year. Here’s a photo I took of a snack stop (no judgement about the food selection) Lets ask ChatGPT about it: What are all the objects in this photo? The photo includes: [Pretty good except for the chips Flavor and it misread longhouse] Are you…