Jaygles

LLMs can only produce evidence. Producing proof requires a human

Two types of devs: Those who require proof before shipping and those who don't

Devs who require proof are shipping high quality. Devs who don't are shipping slop

One-shotting changes without checking them is highly irresponsible

I see people throw around these benchmark numbers

I've seen different benchmarks come to different conclusions

Benchmarking these models must be an incredibly complex and difficult problem

How can a lay person know which benchmarks actually have good signal?

And there is often no respect for serious topics.

Tragedy? Low-effort memes

Someone in high esteem passes away unexpectedly? Low-effort memes

Slow descend into fascism? Low-effort memes

I'd be for forcing companies to list the top-level "owning" entity on all product labels

What downsides have you experienced?

I've been overweight many times in my life, and I've been able to use diet and exercise to get my weight back to a healthy level.

Each occasion took a large amount of willpower and the healthy habits didn't stick around forever.

I started taking tirzepatide weekly and the most striking effect was a marked decrease in cravings. Where previously it took all my willpower to ignore cravings, it has now become difficult to overeat. I would have to force myself.

Now I wonder if its like this for those who have never had issues with their weight. Are there people who insist it's easy but have never experienced cravings?

If your job is to write UI frameworks, by all means do that

Otherwise, yes, you do need React (or something like it). Don't waste time re-solving the problems these frameworks were created to address. Spend your time creating value for your users. Building bespoke state management systems is not needed.

Product labels should prominently display the parent corporation. Whatever is the top of the chain of ownership.

LLM technology will never achieve 100% accuracy in its output. There is an inherent non-determinism. Tasks that require 100% accuracy cannot be handled by LLMs alone. If an LLM is used to replace HR, it will inevitably do something wrong, and a human will need to be in the loop to correct it.

Same goes for chess, there will always be a chance that it makes an illegal move. Same goes for code, there will always be a chance that it produces the wrong code.

Maybe a new AI technology will be developed that doesn't have the innate non-determinism, but we don't have that now.

> The enshitification of the internet is largely driven by people ad blocking

This is unfairly putting the blame on only one rational actor in a prisoner's dilemma.

Content providers are free to put their content behind a paywall with no ads, but they choose not to.

They choose not to because people don't pay for content when they can get it from other providers who don't use a paywall.

Consumers then are left without the option to pay for an ad-free experience.

But ads are run on hardware the consumer owns, consuming their resources and harvesting personal information on the consumer, which is a security concern.

So even if they want to support content creators by viewing the ads they run, they need to also accept the security trade-off, which many reasonably do not

I do want styles tightly coupled to my React components. The product I work on has tens of thousands of React components.

I don't want to have to update some random CSS file to change one component's appearance. I've had to do this before and every time its a huge pain to not affect dozens of random other components. Other engineers encounter the same challenge and write poor CSS to deal with it. This compounds over time and becomes a huge mess.

Having a robust design system that enables the composition of complicated UIs without the need for much customization is the way.

Those technologies don't just solve tech issues, they solve organizational issues. If one or two people manage a website, going without fancy tooling is completely fine. When 1000 people are managing a product with complex business logic across multiple platforms, you need fancy tooling to ensure everyone can work at a reasonable level of productivity.

> No one ever forced any company to work with China.

"Forced" is a strong word here, but company's do need to compete or die. If your competitors are manufacturing in China and selling widgets at a price less than what an American factory can produce them for, what choices do they realistically have?

To expect merchants to get together and act according to some greater good is a pipe dream. Government should have stepped in and prevented the offshoring of American industry through policy

If services offered a paid version that guaranteed privacy, such that I stay anonymous and only data points that are strictly necessary to provide the service are persisted in the company's servers, I would happily pay.

And I mean guaranteed in a way that I would have legal recourse against the company if they go back on their word or screw up