Thread by @SteveSokolowsk2

1) Today's #GPT4 thoughts after another day of use:

People are worried about AI "automating" the writing of code, causing developers to be fired. The reality is that as of last week, the world shifted to where efficiency and computing cost is now the only reason to write code.

2) You can pretty much ask #GPT4 to execute any code you can think of, because natural language is essentially code execution. Last night, my brother used @turbotax for part of his taxes, and asked #GPT4 to complete the theft loss forms for the @GenesisTrading scam.

3) #GPT4 provided accurate information and was able to explain exactly what to do, and it knew that @CelsiusNetwork could be treated as a Ponzi scheme and how the IRS has a specific document for Ponzi schemes as a result of the Madoff scandal.

4) However, #GPT4 used a lot more computer power to generate the output than @turbotax did. It is much more expensive and slower to use a huge neural network to compute the data for the forms than it is to use a rules-based program like TurboTax.

5) But in this case, speed doesn't matter, because @turbotax only performs simple math. So as soon as it is possible to execute more than 25 queries in 3 hours, and get the words back faster, all software where performance is unimportant - like TurboTax - will be obsolete.

6) This realization leads to two conclusions. First, a significant amount of software currently on computers is already obsolete. We don't need #GPT4 to help developers write faster code; we just don't need a lot of rules-based software anymore at all.

7) Second, as I pointed out yesterday, even at the current level of development, the #1 issue limiting the use of general AI like #GPT4 (and it is general AI) is simply that there aren't enough AI chips in the world that can be produced cheaply enough.

8) Therefore, my choice is to short firms that program rules-based software that inputs and stores simple data, like accounting products and POS systems. It's also to buy chip manufacturers that have spare fabs because there is infinite demand for chips.

9) So to summarize: when looking at software, ask: is the software computationally simple, or is it computationally complex? If it's simple, the AI will just do it. These AIs won't be emulating 3D videogames for some time, but they will write code to run them efficiently.