fire_lake

Most developers use languages that lack expressivity. LLMs allow them to generate the text faster, bringing it closer to the speed of thought.

Which also fits with how it performs at software engineering (in my experience). Great at boilerplate code, tests, simple tutorials, common puzzles but bad at novel and complex things.

Please move out of the territory of grammar to make room for AI

The problem is potential residents don’t get a say - only the incumbents. The only solution is national level housing policy.

Users expect near-native responsiveness, even when offline

Do they? I think users are amazed if anything works without internet with modern applications.

There is an F# implementation (Microsoft flavoured ML) called DiffSharp

The post developer era will be the post white collar era.

Surely the level of intellectual difficulty in software engineering is similar to practicing law, medicine, banking… ?

For example, navigating the complexities of legacy systems and workflows and human interactions needed to keep things working.

Well this sounds delightful! Glad to be free of the thinking and creativity!

TDD is opposed to suffering orientated programming IMO.

What pain does writing the tests first resolve? Tests should only be added when you are scared of making changes or want to support assumptions about how things work.

I have seen this first hand with Bazel. You have lots of Bazel rules that are partial reimplementations of the language specific tooling. It usually works better - until you hit a feature that isn’t supported.

Probably means a lack of out of the box rulesets. Writing your own rules to use a build system for a typical project is an unrealistic ask, IMO.

Not a search engineer, but wouldn’t a cache lookup to a previous LLM result be faster than a conventional free text search over the indexed websites? Seems like this could save money whilst delivering better results?

Massive search overlap though - and some questions (like the golf ball puzzle) can be cached for a long time.

This is too simple. Prices float somewhere between cost to produce and value to consumers.

You can’t raise prices beyond value or no one will buy it.

You can’t lower prices below costs (for too long) or you go out of business.

Competition pushes prices down towards costs.

Therefore businesses are always looking for markets with barriers so they can rise prices to value.

Tariffs typically raise costs for all producers, but this only inevitably leads to price increases when competition has driven prices down to near costs.

Unfortunately many staple grocery products fall into this category.

What if people are manipulated by bot farms and think tanks and talking points supported by those corporations?

I think this view of humans - that they look at all the available information and then make calm decisions in their own interests - is simply wrong. We are manipulated all the damn time. I struggle to go to the supermarket without buying excess sugar. The biggest corporations in the world grew fat off showing us products to impulse buy before our more rational brain functions could stop us. We are not a little pilot in a meat vessel.

If you genuinely believe this, why on earth would you work for OpenAI etc even in safety / alignment?

The only response in my view is to ban technology (like in Dune) or engage in acts of terror Unabomber style.

OpenBrain still keeps its human engineers on staff, because they have complementary skills needed to manage the teams of Agent-3 copies

Yeah, sure they do.

Everyone seems to think AI will take someone else’s jobs!

Are you less productive though? Businesses ultimately only care about what gets shipped.

Pythons approach to variable declarations makes this doubly bad.

How so?

Well I don’t know what issue you have with whitespace, but the usual complaint is that programs with small typos are syntactically valid but logically incorrect. This is why CoffeeScript became so disliked. A static type checker makes this scenario much less likely since it won’t compile.

I would also add that F# has some indentation rules (“offside rules”) and tabs are disallowed, further shrinking the input space.

Static typing removes the downside of whitespace.

Oh, and every language with line comments (so most of them) has significant whitespace.

People really like their early return statements however.

And they still haven’t! Although expressions over statements would be a breaking change - I can’t see any mainstream language making this switch.

Old buildings are often stronger because they didn’t have the modelling and precision manufacturing required to make them to an exact standard - instead they were overbuilt.

Guaranteed TCO is pretty rare unfortunately.

Java, Go, JavaScript all lack it.

If the future is not LLMs, then this is a waste of time.

If the future is only LLMs then we’re all cooked.

If the future is a hybrid, then those “grey beard” skills are a much higher barrier to entry than a few months tinkering with ChatGPT.

Please stop spreading FUD.

For this to succeed I think we need a C implementation that can be used by many languages and runtimes.

Lightning is still better than USB C in terms of physical connector design (Lightning puts the male part on the more easily replaced cable side). Annoying that it’s not a strict improvement being imposed.

And even then, execution depends on your language and compiler.

Write a recursive BFS in Haskell and it won’t blow up the stack.

Yes it’s much worse. With Wikipedia we all see the same output and can review it together.