mattvr

Anyone have routing harnesses like this describes with Claude Code? Or other good routing platform recommendations?

(yes, I know this article is about an oracle router)

Is there information on the license of this flag? Not how/if clear you can reproduce it legally.

This helps explain the dissonance in Microsoft's recent Humanist Superintelligence article about creating "AI companions for everyone".

https://microsoft.ai/news/towards-humanist-superintelligence...

In an otherwise pleasant, humanist framing, they jarringly conclude Microsoft's primary AI application will be putting people into parasocial AI relationships for profit.

Ah yes, viral AI gain-of-function research in a secure lab. What could go wrong?

It implies high performance, reliability, and a higher degree of mastery of the developer.

(Which may not all be true, but perhaps moreso than your average project)

Deno Deploy is actually an excellent product.

My choice ranking is Deno Deploy > Fly.io > AWS for new projects, depending on complexity and needs. They also have a new Deno sandbox feature which is great for running untrusted code, AI agents, etc.

The real question is can they adapt to customer feedback fast enough, focus priorities, adequately market & grow, make it profitable, etc. Bumpy road but definitely not doomed.

[0] https://deno.com/deploy

In a nutshell, ideally `peer.connect(remoteId)`. An API like peer-js/simple-peer. And symmetric negotiation would be great as well.

WebRTC should be the universal networking primitive for the next phase of the web, but the API exposes too many implementation details – its abstractions leak.

This plus the overall weight of integrations limit mass adoption by developers.

While WebTransport is promising, it's limited to client-server communication unlike WebRTC.

WebRTC supports peer-to-peer UDP connections as well. Thus it's better for use cases like low-latency games, video calling, and secure direct communication between devices.

A better push might be to make WebRTC more simple and modern, but I'm not sure if any standards committees are working on this yet.

I believe this article is largely wrong and misleading.

Pattern matching is still Stage 1, meaning it’s not a standard: https://github.com/tc39/proposal-pattern-matching

Pipeline operator is Stage 2 and won’t use the “|>” syntax: https://github.com/tc39/proposal-pipeline-operator

The detectors are wrong. Here’s the thing: AI slop has a distinctive structure that many of us spot from a mile away.

The kicker? This setup-punchline format sets off a red alert for astute readers’ AI detectors.

This isn’t just AI slop, it’s an industrial AI sludge factory.

(note: this was ironically written by a human)

You could also use JSON Merge Patch (RFC 7396) for a similar use case.

(The downside of JSON Merge Patch is it doesn't support concatenating string values, so you must send a value like `{"msg": "Hello World"}` as one message, you can't join `{"msg": "Hello"}` with `{"msg": " World")`.)

[1] https://github.com/pierreinglebert/json-merge-patch

Many mobile apps encounter this because React Native still doesn't have a good solution for selectable text [0].

Workarounds exist [1], but aren't great for text that spans multiple lines and styles.

[0] https://github.com/facebook/react-native/issues/13938

[1] https://github.com/bluesky-social/react-native-uitextview

It does appear to be supported by iOS Safari 18.2+, which is the majority of iOS users.

https://caniuse.com/?search=view%20transitions

At Case Bonita in Denver, you have a flag on your table you can raise to signal to the waiter to come over.

This exists: jsonc – and it's somewhat widely used, such as for VS Code configuration.

Also check out the unbranded version: https://www.youtube.com/watch?v=UqJJktxCY9U

Should order this list by number of affected rather than alphabetical IMO. The 275K monthly user platform is almost hidden relative to the 49 and 300 user examples.

Excellent idea! I'd focus on the open-sourcing and community elements (e.g. creating feeds) so it can get traction, rather than features like Markdown and longer posts at this stage.

Hi real quantum physicist, yup, you pretty much got it. This is a prototype demo of the concept – and is using real IBM quantum computers. The queues are between 10s and 10min generally.

This is explained very closely to what you've said in the "Technical details" help section. Occasionally you'll get a real-time result. I'll check out the links you sent next, thank you!

Do you think qBraid could support this with more real-time latency?

Sorry about that. I toned down the animation and fixed the infinite spin issue – curious if it's more comfortable now.

Updated to "favors one side 51%".

You are precisely right.

Yes, sorry, server was a bit overloaded. It should be more reliable now.

Hey HN! Creator here. Sorry for the downtime and dizzying spinning coins.

I was surprised to see this on the frontpage this morning and the scale is pushing the limits of our quantum randomness generator

It should be working again now as I'm pushing fixes. Thanks for your patience.

You could also work around this by adding a scaling transformation that normalizes and centers (e.g. sklearn StandardScaler) in between the raw embeddings — based on some example data points from your data set. Might introduce some bias, but I’ve found this helpful in some cases with off the shelf embeddings.

How about showing all the money you spent/earned in the previous day, in one personalized receipt?

Hey, thanks for the critique (this is my product)! Yeah, LLM summarization definitely has its challenges. We've actually spent months optimizing for this:

1. Using latest models (gpt-4o & claude-3.5) which show major improvements in quality

2. Extensive prompt engineering focused on distilling surprising insights vs. pure summarization

3. Structured output format that's easier to parse

Re: transparency - you're right, we need to add more details. I'm new to this and still learning how to communicate this.

Btw, it's free to use 1x/day, and here's an example: https://nuggetize.com/n/claude-shannon-wikipedia-47ca5e5b

If you're interested in this, you might like https://nuggetize.com/ which works without installing anything and with any browser.

Looks much better!

I could definitely see using this, but don't think it solves anything for me right now.

I think if the diagrams didn't require horizontal panning, and/or supported interactive collapsing, it would be more effective. As it stands, the text view on the left side of the screen is actually easier to parse at a glance than the tree.

Generally, vertical scrolling is more natural for reading this kind of data, and also works better on mobile.

So a first guess of an improvement is something like: the tree diagram generally grows flat & vertical, with horizontal panning (or a manual "expand" click) reserved for deeply nested objects.

The example on the site is a very simple JSON object that doesn't demonstrate JSONtr.ee's potential at all.

To showcase it better, I'd like to see more complex objects on the initial reveal, so I don't have to go and find & format a JSON object just to see if the app/library is worth using yet!