> The right way might be to fight AI slop with AI enforced guard rails.

Whenever I you tried to develop using guardrails with LLMs, I found out that they are much better at ,,cheating'' than a human: getting around the guardrails by creating the ugliest hacks around them.

Mostly works for me. Most of my projects I have some guardrails for what to do around testing, deploying, etc. Seems to work. You are right that LLMs are good at avoiding work and finding loopholes to do so. But generally if you ask codex to "hey look at my gh prs and label the ones that don't meet the contributor guidelines with 'slop'" it might do a decent enough job. Maybe add a skill that spells out criteria. Maybe set up openclaw or similar to do this every morning and then give you the list of prs it will auto close after you say the word.

There are even simpler things: the rating system. There's no guarantee that the driver won't see what I rated him, so I won't report them.

There are ways to report if a man has been sexual with a woman, but they somehow just don't get kicked out of the driver network.

Also just a simple example: Uber engineering blog is full of examples of how they rewrote their app in native Android then web then native again, but nothing about how to solve the real problems humans experience when driving with them.

It just feels that they view Uber as a simple logistic problem where drivers / riders are interchangeable and less like Tinder that tries to match people with similar scores abd kicks out the worst.

The main problem I see is adding things slowly instead of automatic rewrites.

I remember adding lifetimes in some structs and then wanted to use generics and self pointing with lifetimes because that made sense, and then it didn't work because the composition of some features was not yet part of Rust.

Another thing: there are annotations for lifetimes in function signatures, but not inside the functions where there is a lot of magic happening that makes understanding them and working with them really hard: after finally the borrow checking gave me errors, that's when I just started to getting lots of lifetime errors, which were not shown before.

Rust should add these features but take out the old ones with guaranteed automatic update path.

The edition mechanism covers your last paragraph.

It's a product category name at least, not a release name, so the next release can be Neo 2

This is just not true, people get promoted for delivering impact whether the solution is complex or simple.

The best engineer I know who can work with huge complex systems in a big company usually starts with a complex solution then after he understands what he wants to achieve thinks backwards and reimplements it in the fewest possible lines of code change with the already complex system.

There's exception and geniuses to every rule. In general however a simple solution will be much more difficult to argue a promotion around even if you make a ton of impact. You may get a top rating and a slightly larger bonus however not a promotion.

Every large company has a ladder for promotions that includes many words that basically come down to "complex." "Drive a year long initiative" or "multiple teams" or "large complex task with multiple components" are all examples I've seen.

Yeah that large company promo thing drives me nuts. Perpetual gaslighting "meh... that was too easy lel". Yeah thanks. Often stuff isn't easy but hard to explain why if the solution turned out to look easy.

What is funny is you can dance through the hoops for 3-5 years for promo. Or grind leet for 100 hrs and get it by jumping.

Was going to say something like this. If you're good at keeping things simple, it will help you deliver impact which can get you promoted.

I’m here to support both of your statements. This is absolutely true from orgs the size of FAANG to startups because I’ve worked at both. Sure smooth talkers get promoted but so do smart people who make things work better by simplifying.

The main lesson is quite simple: if you can write the test to be uncheatable, ChatGPT can write the code for it.

Is there a way to get sticky model selection back, or the reason is that it is just too expensive to serve alternative models?

For coding I love codex-5.3-xhigh, but for non-coding prompts I still far prefer o3 even if it's considered a legacy model.

I can imagine that its higher tool use is too expensive to serve, but as a pro user I would love it to come back.

Sure, and actually the open models are already good enough to do that, it's not like any company could stop any organization that can collect the data from doing this.

They can just improve on it a lot.

I don't really understand this reasoning actually:

if OpenClaw usage go up, and a service (OpenAI it looks like) gets lots of usage data for personal assistent usage, they can optimize to make it better for people who get a $200 subscription just because of that use case.

For anybody who thinks it's about Trump vs other administration: it's not, both AI surveillance of all people and using it for automatic fight was just bound to happen.

The only question is whether the safety of the models were really done well enough to protect the people and be a net positive force in the world.

I guess if they would be safely trained to do more good than bad (how Dario and SamA said), there wouldn't even be a need for the contract terms.

It would/will be extremely irresponsible to put non-deterministic and fallible models in charge of weapons. We are not close to having solved the problem of ensuring AI pursues good outcomes

I agree completely. Anybody who uses the models extensively know it can do something amazing for a prompt and something awful for another. But I also know that wars are unfortunately real and there are real enemies between countries and they don't want a limited model.

How exactly does the "limitation" affect any war the US may be in with another country?

Probably drones targeting and automatically killing Russian people by a thinking model guessing if its Russian on Ukrainian person is a red line.

Elon Musk already denied Starlink for being used for remote killing, but at some point all these technologies will be nationalized, as they are too important not to be.