btown

21 ms·

421.
by btown 1y ago
As a VS Code + Claude Code user, I'm really excited to see progress here, because the official near-zero-config Claude Code IDE integration is... inflexible, at best. What if I want to send a subset of my open editor panes to Claude Co
422.
by btown 1y ago
I have a bunch of points. I want to fit a curve to them. I could write a function that takes a bunch of parameters as floats that specify the curve, and an x coordinate as a float, and have it output the most likely y value as a float. If I
423.
by btown 1y ago
If you store a specification for a probability distribution and give that specification a name, it can behave like a function in that it can be sampled for a scalar output. It can behave like a variable in that you can assign it to a new va
424.
by btown 1y ago
https://colcarroll.github.io/ppl-api/ is likely a good starting point to get a taste of examples in Python; some use custom languages, but the success of Python-native frameworks in the LLM world I think has shown that
425.
by btown 1y ago
Once one understands that a variable (in a programming context) can hold a specification for a variable (in a mathematical context), one opens up incredible doors that are at the foundation of modern AI. When you see y = m * x + b, your r
426.
by btown 1y ago
This makes me wonder: are there "cloud drive virtual sneakernet" systems that will communicate e.g. by a client uploading URL request(s) as documents via OneDrive/SharePoint/Google Drive/Baidu etc., a server reactin
427.
by btown 1y ago
It’s unclear how that phrase got into the wiki; the bill text https://billstatus.ls.state.ms.us/documents/2024/html/HB/110... does not seem to mention that phrase.
428.
by btown 1y ago
I'd go a step beyond this (excellent) post and posit that one incredibly valuable characteristic of traditional NLP is that it is largely immune to prompt injection attacks. Especially as LLMs continue to be better tuned to follow inst
429.
by btown 1y ago
As someone who's been saved by look-aheads in many a situation, I'm quite partial to the approach detailed in [0]: use a regex library that checks for a timeout in its main matching loop. This lets you have full backwards compatib
430.
by btown 1y ago
"Just use Nginx" was not a viable option here, without additional Certbot etc. orchestration, until 14 days ago! And this is still in preview! https://blog.nginx.org/blog/native-support-for-acme-protocol And,
431.
by btown 1y ago
Sure, there is an argument about slippery slopes here. But the thing about the adage of "if you slowly boil a frog..." ( https://en.wikipedia.org/wiki/Boiling_frog ) is that not only is the biological metaphor
432.
by btown 1y ago
Don't worry! If ink's too expensive to buy outright, have no fear - you can buy a subscription to Instant Ink and pay... (checks notes) even more! But you'll never need to worry about buying ink again! https://ww
433.
by btown 1y ago
> without even looking at that author's past It is worth noting that this comes two weeks after the authors posted https://arxiv.org/abs/2507.12213 "Is the Interstellar Object 3I/ATLAS Alien Technolog
434.
by btown 1y ago
There's some nuance here. The compiler flags SQLITE_DEFAULT_SYNCHRONOUS and SQLITE_DEFAULT_WAL_SYNCHRONOUS are set to FULL by default, which does fsync on each commit. https://sqlite.org/compile.html But there is a thi
435.
by btown 1y ago
Having spent a lot of time lurking on the frustratingly-slow-moving bikeshedding thread for the Javascript pipe operator [0], there's a great irony that a lot of people want a pipe operator because they don't want to deal with f
436.
by btown 1y ago
Sadly many things define the __or__ operator, including dicts and sets which are common to find in pipelines. ( https://peps.python.org/pep-0584/ was a happy day for everyone but me!) In practice, rshift gives a lot mor
437.
by btown 1y ago
My favorite Python "crime" is that a class that defines __rrshift__, instantiated and used as a right-hand-side, lets you have a pipe operator, regardless of the left-hand-side (as long as it doesn't define __rshift__). It&#x
438.
I Have No Mut and I Must Borrow(old.reddit.com)
7 points by btown 1y ago|0 comments
439.
by btown 1y ago
One of my mental models is that the notion of "effective engineer" used to mean "effective software developer" whether or not they were good at system design. Now, an "effective engineer" can be a less battle-t
440.
by btown 1y ago
I'm really surprised to hear this. If anything, I'd expect that every company transitioning from > "we want to store/retrieve thin event logs and clickstreams" to > "we need to store/retrieve/jo
441.
by btown 1y ago
As a non-lawyer who’s nonetheless been asked to help to review internal documents en masse - the idea of a fully scriptable <50ms switch time between documents is quite appealing. AI can help with initial screening, but there are many si
442.
by btown 1y ago
I think there’s a really interesting point here that even if a model is capable of planning and reasoning, part of the skillset of a creator is that asking for the right thing requires an understanding of how that model is creating its arti
443.
by btown 1y ago
You’ll probably hear people say things like “this could just be an MCP server and a prompt to use it.” To that I’d say: just remember that infamous 2007 Dropbox comment: https://news.ycombinator.com/item?id=9224 If you can
444.
by btown 1y ago
It's an interesting philosophical question. Imagine an oracle that could judge/decide, with human levels of intelligence, how relevant a given memory or piece of information is to any given situation, and that could verbosely desc
445.
by btown 1y ago
I can't give advice about the trademark, other than that I've been through this kind of thing before and it sucks . What I can say... is that I love what you've been doing on your Deepkit, and I was horrified to hear that
446.
by btown 1y ago
Simple: An LLM can't leak data if it's already deleted it! taps-head-meme
447.
by btown 1y ago
A "yellow flag" moment for me was OpenAI's revision of their Preparedness Framework in April 2025 to remove "risks related to persuasion" from the framework altogether. https://cdn.openai.com/pdf&#
448.
by btown 1y ago
This makes me really want to know what would happen if you took the entirety of Dan Abramov's github comment history and transposed it to the style of Why's https://poignant.guide/ .
449.
by btown 1y ago
These aren’t third-party libraries, these are RFP processes for adding to the official protocol, and standardizing the semantics of new fields and data types. A world of difference IMO.
450.
by btown 1y ago
If you want the things mentioned in this article, I highly recommend looking at https://github.com/modelcontextprotocol/modelcontextprotocol... and https://modelcontextprotocol.io/community/sep-gui

More ›