We built computers to interface with us in our language. In my case that’s English.
We taught computers statistical English to make AI/LLMs.
Some AI things are a reminder how simple humans can be.
Your vocabulary may only be 40,000 words.
I understand the frustration from PocketOS and the deletion event that happened with Claude & Railway. However, the glaring gap for me, was them not having a basic constructed BCDR that included off-platform/off-site/airgap backups.
Agents will need sick days. They will need to self report an understanding of their hallucination episodes and be removed from their queue of work. Immediately triggering a removal from their Web3 payment plan.
I’m currently only interested in multi agent offline machine psychology. I need multiple agents intentionally running different LLMs. The reminder here is there is no perfect LLM, or winner. You need the inherit flaws in all of them to run a diverse agent company.
We’ve known all along that LLMs are statistical language pattern matching machines. People talking about that Apple paper like it’s new knowledge. An LLM isn’t knowledge, it’s an ADHD engine that still does misremembering.
An increase in security issues across the Internet will be met with a proportional response to increase data sovereignty. Risks will change. Surface areas will change. The pendulum has always swung in both directions.
@0xTib3rius Lawyers that learn and use certain aspects of AI, like in research, may eventually make them outshine lawyers that do not use AI. Lawyers still have in person and in court demands that will keep them professionally employed basically forever. They are not at risk from the system.
NVIDIA CEO Jensen Huang: “I really discourage 1-on-1s”
Jensen famously has 60 direct reports. When Stripe founder Patrick Collison points out that this isn’t conventionally considered best practice, Jensen shares his reasoning:
“I don’t do 1-on-1s, and almost everything I say, I say to everybody all the time. I don’t really believe there’s any information that I operate on that only one or two people should hear about… I believe that when you give everybody equal access to information, that empowers people. And so that’s number one… Number two, if the CEO’s direct staff is 60 people, the number of layers you’ve removed in a company is probably something like seven.”
Patrick offers to steal man the other side of the argument:
“1-on-1s are where you provide coaching, where you maybe talk through personal goals and career advancement, where maybe you give feedback on something that you see somebody systematically not doing so well… Do you not do those things or do you do them in a different way?”
Jensen responds:
“I give you feedback right there in front of everybody. In fact, this is a really big deal. First of all, feedback is learning. For what reason are you the only person who should learn this?… We should all learn from that opportunity… Half the time I’m not right, but for me to reason through it in front of everybody helps everybody learn how to reason through it. The problem I have with 1-on-1s and taking feedback aside is you deprive a whole bunch of people that same learning. Learning from other people’s mistakes is the best way to learn.”
Video source: @stripe (2024)
@SwiftOnSecurity I totally agree. GPOs can do an incredible amount of lifting. I’ve fixed massive organizations with GPO inheritance. One thing that usually has to go with it is OU design and permissions. Dealing in blocks is harder than getting the structure right.