smallnix

Yes https://physicsworld.com/a/dark-coated-starlink-satellites-a...

Could use some non-linear transformation to not be "stuck" too long on 1% and 99%. I.e. reaching 2% in ui while actual progress is still below 0.01.

Maybe fast in the beginning and end and slow in the middle, "ease-in-out" or the inverse with sigmoid

Brave to make ads with the Ghibli style. Would have thought that's burned by now.

Awesome project. Gutenberg is mentioned, does this project feed back to Gutenberg?

Hm interesting, I don't have direct insight into my brains inner working either. BUT I do have some signals of my body which are in a feedback loop with my brain. Like my heartbeat or me getting sweaty.

Nice analysis. Google should take notice. Do worldwide used apps do this too?

> Apple’s obsession with privacy and security is the reason most of us don’t think twice to scan our faces, store bank account information or share our real-time location via our phones.

People do that because it's very useful, not because it's safe.

*stupid serial killers

+1 matches my experience, https://en.m.wikipedia.org/wiki/Conway%27s_law

You don't. You don't pay an AI. You pay a company owned by humans that offer an AI service.

Which companies did the author start/build?

> Why should one feel guilt about being prosperous in an unequal society?

I can understand the idea of feeling guilty about wasted potential (wealth, time, strength, beauty, intelligence). That which could be used to help those who need help, not exactly novel: “If you have two coats, give one away”

Thanks for the blog post, I like these personal pieces of software history

Is my understanding wrong that LLMs are trained to emulate observed human behavior in their training data?

From that follows that LLMs fit to produce all kinds of human biases. Like preferring the first choice out of many, and the last our of many (primacy biases). Funnily the LLM might replicate the biases slightly wrong and by doing so produce new derived biases.

So it's an indirection layer to the actual APIs.

The value of MCP then depends on it's adoption. If I need to write an MCP adapter for everything, it's value is little. If everyone (API owners, OS, Clouds, ...) puts in the work to have an MCP compatible interface it's valuable.

In a world where I need to build my own X-to-USB dongle for every device myself, I wouldn't use USB, to stay with the articles analogy.

Found one https://arxiv.org/abs/2404.06654

Do I understand right this requires access to internals of the LLM and can not be used with todays models behind an API like ChatGPT or Claude?

In java usually clazz

I am making extensive use of deadlined passive approval for decision I consider non-contentious. For everything else I set up meetings. I don't believe in async review of contentious decisions.

When serious money is on the line the previously hard rules can become soft quite fast. 1 major escalation away in my experience.

I heard they have strong Nordic developers waiting for free markets in Greenland.

> it only gains features, never loses them

The article is about losing features due to incompatibility of plugins. Perhaps you refer to some more abstract open source concept?

Are there papers on this effect? That quality of responses diminishes with very large inputs I mean. I observed the same.

I understand the interest about this bug, but to my understanding this is an unpaid hobby project?

If that's true I don't feel entitled to expect anything here.

What specifically do you want to protect?

Thanks for sharing your code.

Don't be disheartened by negative unconstructive comments.

I think maybe adding a disclaimer that you use an LLM since English is not your first language will preempt such comments in the future.

Keep at it, I think such tools are great to build, share and learn from.

Keep focusing on functionality!

> For a century, the business of running a Hollywood studio was straightforward. The more people watched films, the more money the studios made.

I thought Hollywood (Disney) long before Netflix tapped into other revenue such as merchandise.

I think it's ok if e.g. people with a different first language than English use LLMs as some sort of "smart spellchecker".

That the software does some error handling while performing it's goal is not a red flag?

Is the usecase to /1 "make symlink" clones that can be organized in a different way or to /2 move the originals somewhere else and replace with symlinks?

Loneliness is longing for connection to other people.

Loneliness does not need to relate to the number of people you talk to or that you know.