HN user

cwilby

52 karma
Posts5
Comments49
View on HN

This is it right here. I love the Python ecosystem but if I can find a JS alternative I'm using that for ergonomics.

Is it just me or am I not seeing any example that isn't pure theory?

And if it is just me, fine I'll jump in - they should also make it so that users have to approve local network access three times. I worry about the theoretical security implications that come after they only approve local network access once.

It's 2025's ROT13 cipher for API. /s

It's also "Model Context Protocol", a protocol for LLMs to interact with third-party services.

Internally, we call the request’s destination server its “origin”

Origin: The point at which something comes into existence or from which it derives or is derived.

How can the request's destination server be the origin, if it is the destination server?

GitHub was down 2 years ago

It looks to be contained to just GitHub, azure service page shows no outages at this time.

This tracks - it's difficult to speak at the same pace in Chinese as I can English. That said - are those 39.15 bits plaintext? Compressed? Encrypted?

The size of a word does not correlate with it's concept - I still posit that some languages can transfer concepts faster than others, minus our baud rate.

Edit: Or, perhaps I am not as gifted an English speaker as my bias has presumed :| For example, I had to lookup "syntagmatic".

I just started learning Chinese about 2 months ago, to me it seems they stuff whole concepts into characters.

For example,

"去" (pronounced "Qú") is "going to the". "超市" (prounced "Chao Shi") is "supermarket" "去超市" (pronounced "Qú Chao Shi") is "going to the supermarket".

3 syllables vs 7 syllables.

To me, it seems that instead of composing letters into words to convey meaning, they have more letters that are mini-words unto themselves.

I don't know what to say more than - I wish they'd calculated in KWh instead of $.

I really want to know that figure, but I'm left extrapolating.

I could _guess_ it's huge, but that's not the point.

"Can we make a separate intelligence beside ourselves, damn the energy expenditure?"

RoR Debugbar 2 years ago

We recently installed something similar (https://miniprofiler.com) into a legacy ASP.NET MVC app. It is setup to only be visible in local development environments.

In areas of the code where ORMs obscure the actual SQL that runs, it's shortened the amount of time and effort it takes to discover a slow route and optimize it.

"Handy to have that extra info in your dev environment" is pretty spot on.

It is interesting to me that the approach required to work with this tool, is almost identical to using every other tool.

It boils down to - "try breaking this problem into smaller problems to increase the solution space".

I kind of agree and disagree (enough to type some stuff)

I can see how interaction with a human-like device is more engaging than Alexa (I have a Google Home speaker and feel the same)

However, then we're entering "uncanny valley", where we're always acutely aware that we're not talking to the caliber of consciousness we're used to.

Other companies have tried this: https://www.engadget.com/2018-07-25-adorable-home-robot-kuri...

Whatever voice assistant "dominates" next should be fun and not get in the way. Two very complex requirements in that particular domain.

Couple of clues

The author mentioned the tool sends out requests, and just measures response time regardless of response status. (If they fail, they fail!)

And they measure response times from gstatic.com, which should have edge nodes near most locations.

Shot in the dark guess - perhaps you're hitting some kind of rate limit somewhere in that chain?