Social media felt this way in the early days, too.

> they almost never feed me misinformation (and never deliberately)

This is dangerously untrue. LLMs will deliberately tell you the things you want to hear. They even "call you out" in exactly the ways that are most palatable to you. There is ample evidence that LLMs can be biased. The ability to broadly influence individual opinions, without even needing bot accounts on social media is exactly one of the reasons so many people who want power/money are flocking to control these LLMs.

> so many of my interactions are with jerks

This is one of the ways that the LLMs are "scarier" and more powerful than social networks. Real people are more likely to be obviously pushing agendas or otherwised easily dismissed.

> LLMs are the ones getting better

Getting better at what? The things you list are exactly the things that make LLMS subtly powerful and far more dangerous that social media. It's the reach and anonymity of social media without needing to control a feed algorithm, deal with content or people that doesn't fit your agenda, or produce fake content that does.

If you actually want to understand why "an actor pretending to enjoy drinking Coca Cola" is okay while other kinds of endorsements are not, in the US at least, you could start with the FTC "Truth in Advertising" website, and "Advertising Endorsements" specifically, or other resources on that page.

https://www.ftc.gov/news-events/topics/truth-advertising

https://www.ftc.gov/news-events/topics/truth-advertising/adv...

If picking real winners and real winnings to feature in the ad was just as good, they could do that. If not, then yes, it makes an impact on the world to mislead people with that marketing.

Somehow there's a difference between things that happened and didn't happen, and that's a good place to draw a line in the sand of what you're allowed to advertise and not.

I think it can be easy to look at code as an asset, but fundamentally it is a liability. Some of the "bottlenecks" to new code are in place to make sure that the yield outweighs the increased liability. Agents that produce more code faster are producing more liability faster. Much of the excitement and much of the skepticism about coding agents is about whether the immediate increased productivity (new features) and even immediate yield (new products or new revenue) outweighs the increased long term liabilities. I'd say we won't find out for another 1-3 years, and of course that the answer will differ in different domains.

From this perspective, attempting to build these bottlenecks into the agentic workflow directly makes some sense. Supplying coding agents with additional context that values a coherent project vision and that pushes back against new features or unconstrained processes would be valuable.

Is this what the article is trying to get at? Is this attempting to make some agents essentially take on product management responsibilities, synthesizing as much as possible into a cohesive product vision and reminding the coding agents of that vision as strictly as possible? Should these agents review new proposals and new pull requests for "adherence to the full picture", whether you want to call this "context" or "vision" or something else?

I think these agents might do an exceptionally good job at synthesizing context and presenting a cohesive roadmap that appears, linguistically, to adhere to the team values and vision. But I'm doubtful that they can have the discernment that a quality manager or team can have. Rapidly and convincingly greenlighting a particular roadmap could do more harm than good.