I wish the author wrote more about the day 2 problem cases with AI built applications. It somewhat matters what the programming language is, the architecture and design for debugging and reasoning verification when we want to alter the system specification.

Basically as a Dev, or "Owner" of the application, we are responsible for the continuous changes and updates to the system. Which I've found hard to reason about in practice when speaking to other people, if I dont know the code explicitly.

In your comparison, I think you're getting tired because you need to reason through syntax writing which you have not done in a while, its more akin to going to gym than it is to "withdrawl". Having to reason through syntax for writing is going to be a muscle memory thing which AI will reduce, but its not really the most valuable skill of programming. Reading, comprehension, ownership, and your internal mental model of programming tasks are without a doubt the most important aspect of our work.

I think the idea raised about "Automated Firms" is a bit off in the picture painted in that linked article. I think the David Oks intention is to paint a picture of a fully automated company, but the linked article gives this impression:

> Future AI firms won’t be constrained by what's scarce or abundant in human skill distributions – they can optimize for whatever abilities are most valuable. Want Jeff Dean-level engineering talent? Cool: once you’ve got one, the marginal copy costs pennies. Need a thousand world-class researchers? Just spin them up. The limiting factor isn't finding or training rare talent – it's just compute.

In that above paragraph the author is saying to the reader that a human will be able to spin up and get these armies of intelligent workers, but at the end of the day their output is given to a human who presumably needs to take ownership of the result. Intelligent workers make bad choices or bad bets, but those AI machines cannot "own" an outcome. The responsibility must fall on a person.

To this end, I think the fully autonomous firm is kind of a fallacy. There needs to be someone who can be sued if anything goes wrong. You're not suing the AI.

1. its apart of an enterprise sales motion, eg you want to get ~1000$ per user per month out of it. 2. Its apart of some well subscribed newsletter, this looks more like an ad than a cold email.

If your business is trying an outbound sales motion, then these methods might work. If you're going down a Product lead growth motion because of the nature of the problem/product/customer, then I would try getting some influencers who your prospective users like to use your product.

I know a lot less about PLG motions, so take that with a grain of salt. Feel free to email me directly if you want feedback or anything. Email is in my profile.

For getting new users, traditional outbound motions like cold email, and cold calling work. But only if you know the problem youre trying to solve really well.

We I 787 I 879-0215 I I I ui 87⁸⁸78⁸877777777 I 77 I⁸7 I 87888887788 I 7788 I I 8 I 8 I 788 I 7⁷88 I 8⁸I 7788 I 787888877788888787 7pm I 87 I⁸77 I ui 77887 I 87787 I 7777888787788787887787877777⁷777⁷879-0215 7777 I 7pm⁷I⁷879-0215 777⁷IIRC 7 7pm 87787777877 I I I⁷⁷7 ui ui 7⁷879-0215 I IIRC 77 ui 777 I 77777 I7777 ui I 7877777778 I7 I 77887 I 87⁷8777⁸8⁷⁷⁸⁸7⁸⁸⁸87⁸⁸⁸⁸8⁷87⁸⁸87888⁷878⁷878887⁸⁸⁸88⁸878888888888888888888887878778788888888787788888888888888888888888888887 ui is 888888888887 7

They still care about the quality of the product, just not the process as much. Not sure if this the case for all people or a generalization. Just something I noticed.

a lawyer and judge are discussing a case and using the terminology CSAM in the case and needs to argue between the legality or issue between the child being real or not. What help is it in this situation to use CP vs CSAM in that moment. I dont really think it changes things at all. In both cases the lawyer and judge would need to still clarify for everyone that "presumably" the person is not real. So an acronym change on this point to me is still not a great take. Its regressive, not progressive.

So maybe on some apps, but "all" is a difficult thing.

Generating pictures of a real child naked is assault. Imagine finding child photos of yourself online naked being passed around. Its extremely unpleasant and its assault.

If you're arguing that generating a "fake child" is somehow significantly different and that you want to split hairs over the CSAM/CP term in that specific case. Its not a great take to be honest, people understand CSAM, actually verifying if its a "real" child or not, is not really relevant.

to be frank though, I think this a better way than all people's thoughts all of the time.

I think the "crowd" of information makes the end output of an LLM worse rather than better. Specifically in our inability to know really what kind of Bias we're dealing with.

Currently to me it feels really muddy knowing how information is biased, beyond just the hallucination and factual incosistencies.

But as far as I can tell, "correctness of the content aside", sometimes frontier LLMs respond like freshman college students, other times they respond with the rigor of a mathematics PHD canidate, and sometimes like a marketing hit piece.

This dataset has a consistency which I think is actually a really useful feature. I agree that having many perspectives in the dataset is good, but as an end user being able to rely on some level of consistency with an AI model is something I really think is missing.

Maybe more succinctly I want frontier LLM's to have a known and specific response style and bias which I can rely on, because there already is a lot of noise.