Roughly agreed. I'm a bit baffled when it seems like someone is having long conversations with multiple tasks and loads of add-ons. I generally have one or two iterations and then a new session.

I'm using another tool, not claude code, but I don't think that matters much.

In many many ways our technology is not working on our behalf. Even though it seems possible.

You might look into applying RL in the domain of low cost robotics and drones. That would draw on some of your past experience but applied to a domain (robotics) which I perceive is seeing renewed interest.

I find value in learning some things deeply but not all things.

The ability to be more selective about where I attend deeply, while leveraging fast shallow learning to complete other tasks... That seems like a potential benefit and a nice choice to have in the toolbox.

trick is maintaining enough domain expertise... so we can actually audit those shallow outputs.

If the baseline knowledge drops too low we cannot tell when the AI is being lazy or wrong

Spot on. The ultimate bottleneck is no longer generation; it's verification.

If you don't intrinsically know what 'right' looks like, AI simply helps you build the wrong thing faster. This internal compass is exactly what I meant by 'taste' in the original post.

It's intriguing how book recs aren't really that good anywhere. I mean they are ok, but rarely great. Google and Amazon have most of the text but still don't solve this so well.

In any event, I sometimes use GoodReads. You can go to the page for a representative book and then start browsing the "readers also enjoyed" list.

I also use LibraryThing in a similar way.

The problem is the perverse incentives of recommendations and profit.

How to solve it?

Why do you think the problem is hard to solve?

It's technically quite feasible with modern search and language model technology.

The challenge is in acquiring all the book text, meaningful cost to implement, and minimal paths to revenue generation.

For various reasons, the organizations that have the text aren't incentivized to deliver best recs for a user, as tim noted.

Significant chunks of the software engineer community self select into minimizing time spent with customers or even any non-engineers. It's understandable but unfortunate too.

I would suggest looking for people that were initially hands on but converted to sales engineer or professional services roles. Some switch because they don't like the programming. Some will quietly be very good engineers that have seen a lot of scenarios.

You'll have to pay a premium for travel if you want to get experienced people that have families at home.

Thanks for the feedback! We're refocusing our efforts to find strong engineers who have the right instincts and soft skills to deal with customers. Travel will still be a part of the role and we're happy to pay for that! I thought this was a hot role, so I'm coming grips to the reality that it is going to be very hard to fill!

- Invest money from each paycheck.

- Make a practice of walking daily.

- Casually lift a few times a week.

- Pray daily. Focus on gratitude.

Invest on what, though.

Basic index etfs for Nasdaq and SP500.

Yes, very realistic.

Search for the engineering / manager pendulum online to see various discussions on the topic.

There are lots of differences between management in a startup vs a larger or more mature company too.

Agreed. You should be able to set the waf to just drop the packets and not even bother with the overhead of a response. I think cloud flare waf calls this "block".

Yeah, this is the way. Dropping the packets makes the requests cheaper to respond to than to make.

The problem with DDoS-attacks is generally the asymmetry, where it requires more resources to deal with the request than to make it. Cute attempts to get back at the attacker with various tarpits generally magnifies this and makes it hit even harder.

When the WAF drops packets, how does pricing work? I am assuming there is still a non-zero cost to handling that? Kind of sounded from OP that they are looking to shake the monkey off their back for good, and cheaply.

Where do you get the data?