metalrain

Even with locally runnable small "open" models you are relying on scraps of others. They are much worse at the LLM game and you don't know when they stop releasing the weights.

How can you go the opposite direction? Instead of using LLMs to produce more code, can you produce less, maybe higher abstraction code?

I think it's appreciation of the world and people to look and think, "some people did that". So many people working together globally to produce anything you see, sometimes over decades and many lives.

There is extraordinary in the ordinary.

In the story they spend months to build the MVP and people don't like it. This clearly is first point where they could "fail fast", but they believe they can improve and they do.

I guess I'm thinking where is the "fail fast" that is fast enough, but also not quitting too early?

Usually no.

And I have to say that no one tries to build a failed business. Founders can be really earnest about their intentions, work harder when they see the cracks, but often it just doesn't work, they don't find the right way before it's too late.

Maybe just before the end someone tries to siphon the funds into private account or assets into their next venture, but you tend to get caught doing so.

Only functional startups I've seen solve actual customer problem and don't try to solve all problems. Usually they have tried several things before they found one that actually is worth solving.

But everyone makes mistakes and bad deals in product development or they go out of business.

Isn't any camera with low enough angle good for license plate reading with additional software?

I noticed that gameplay speed depends on the window size. I'm assuming that larger canvas takes longer to render. It seems too fast at small window sizes and maybe too slow at 4K, not sure what is the intended speed.

I think it's more abstraction problem.

You could write more code, but you also could abstract code more if you know what/how/why.

This same idea abstracts to business, you can perform more service or you can try to provide more value with same amount of work.

That doesn't sound like ban, you have to disclosure yearly the amount of stock you have demolished, but there is no mention of penalty or anything like that.

This is a big deal and it's not just code.

Any decision maker can be cyberbullied/threatened/bribed into submission, LLMs can even try to create movements of real people to push the narrative. They can have unlimited time to produce content, send messages, really wear the target down.

Only defense is to have consensus decision making & deliberate process. Basically make it too difficult, expensive to affect all/majority decision makers.

Communities also evolve and devolve with time even without large external event. Maybe you don't feel the same belonging in the friend group after ten years or community grows to become something it wasn't in the beginning.

Maybe you have to accept that communities are here and now, but they can dissolve at any time.

I think it's the loss of control.

Even if you can achieve awesome things with LLMs you give up the control over tiny details, it's just faster to generate and regenerate until it fits the spec.

But you never quite know how long it takes or how much you have to shave that square peg.

Good point, sales is the winning factor in most cases. Why is Microsoft one of the largest software companies? Sales.

Yes and that is exactly why they are losing. They have hostages not customers.

Yes, many don't like Sharepoint, but still they use it. It's the tool they can use.

Customers don't care if Sharepoint uses LLM, they just want to share ideas, files, reports, pages, etc. If LLM makes it easier, great! If some other product makes it easier, great!

It's not about the product it's about the results.

I see that Software as a Service banked too much on the first S, Software. But really customers want the second S, the Service.

When you sell a service, it's opaque, customer don't really care how it is produced. They want things done for them.

AI isn't killing SaaS, it's shifting it to second S.

Customers don't care how the service is implemented, they care about it's quality, availability, price, etc.

Service providers do care about the first S, software makes servicing so much more scalable. You define the service once and then enable it to happen again and again.

I'm constantly reminded how software is all around us, we don't even notice it.

Operational excellence survives, no matter the origin.

I think we will use more tools to check the programs in the future.

However I don't still believe in vibecoding full programs. There are too many layers in software systems, even when the program core is fully verified, the programmer must know about the other layers.

You are Android app developer, you need to know what phones people commonly use, what kind of performance they have, how the apps are deployed through Google App Store, how to manage wide variety of app versions, how to manage issues when storage is low, network is offline, battery is low and CPU is in lower power state.

Skills.md will in time have same problem as MCP, they will bloat the context. I wonder if we could just have the scripts without the descriptions and LLM would have been trained to search the most useful things in specific folder.

I like the idea but the example doesn't make much sense.

In what application would you load all users into memory from database and then filter them with TypeScript functions? And that is the problem with the otherwise sound idea "Functional core, imperative shell". The shell penetrates the core.

Maybe some filters don't match the way database is laid out, what if you have a lot of users, how do you deal with email batching and error handing?

So you have to write the functional core with the side effect context in mind, for example using query builder or DSL that matches the database conventions. Then weave it with the intricacies of your email sender logic, maybe you want iterator over the right size batches of emails to send at once, can it send multiple batches in parallel?

I think LLMs are best as learning tools, explaining code and producing something that can be then iterated.

You don't need blockchain for that. Total BS.

I think chatGPT is like porn, it suppresses the urge but it doesn't give the resolution.

Neat, but how do you close the message blocking the calendar?

I have found that cooking for family, friends or small communities can scratch that itch.

Results are tangible, you are doing something with your hands, it only takes few hours (often much less) and you get to give something to people you love.

I tend to have very periodical special interests, maybe 1-3 months per topic. It's not long enough to make life changes, to chase that crazy in me.

Maybe I have to become "Today I Learned" style content creator

I've been dating recently and I think planning dates is mostly fantasizing.

I plan the time, place and activities, imagine things to discuss, how I feel, how other person might feel. Everything feels well thought out and detailed, unpacked.

But it never realizes quite like that, plans get cancelled, places are closed, feelings have changed, discussions are sidetracked or don't feel right. It's all revealed to be a fantasy.

I don't mean you shouldn't try to plan, but unpacking can also be fantasy as much as the packed thing is.

I think the problem is that when confronting all future plans you might find out you don't like any of them.

But you have to do something, choose something. So it's almost better if you don't think too hard, just do it and find out, learn to become content with it.

I feel like AI is great tool to expand the environment of your knowledge, kind of like wikipedia (I wonder why ).

But poor as accurate source of knowledge or provider of accurate answers.

People only click from search when they have actual need to visit the site. E-commerce, image/video content, social media, etc.

That is pretty much as it should be.