I'll be impressed when they can create an entire cell from scratch and it will start to divide. They can create all the needed precursors, bypassing millions of years of random permutation. Because until you have an entire working cell with replication, you have no retained benefit.
With all the people saying that you're going to have problems because the LLM is not good at refactoring or large code bases or OOP, etc. the point may be that if you're working to develop skills, an LLM herder might be a good one. Even if the models are almost, but not quite, good enough yet - they will be.
When looking for a career move you might want to focus on the trajectory more than the current state.
this is the pitch - it's open source, run it yourself. But >99% of people will not have the hardware needed to run these models at a high enough quality to be close to SOTA. So they will run the open-source models on CCP systems for a good price.
Unless you're very careful, it's trivial to have my secrets to be sent to the LLM. If it reads your .env just to see the variable names, the secrets have been sent to the servers. Now - they probably don't care about you and your secrets - but it makes me more uncomfortable that they have them.
This is true of anthropic or openai - but for some reason I think the us govt or anyone else will have a harder time getting to my data from them than the CCP will any chinese company.
You'll see a lot of advanced users advising against compact. The truth is that you have the entire transcripts in your ~/.claude/transcripts (I give claude permission to look there) so when there is some important discussion I don't want to lose, I use claude extract the notes from the transcript. If it's a long transaction it can be big so it might miss things on the first go around - but if you ask it specifically about the topic you're interested in it'll usually find it.
29172cd79cb541e2bf2442631a84da5f
0c85ce