senko

32 ms·

661.
by senko 2y ago
Reposted from original: https://crawshaw.io/blog/programming-with-llms HN discussion of original a few days ago: https://news.ycombinator.com/item?id=42617645
662.
by senko 2y ago
I read all the negative/neutral GoodReads comments of the book (and author's other books if I'm not familiar with the author, and maybe Wikipedia if I want to dig deeper). 99% of books I learn about from recommendations (HN,
663.
by senko 2y ago
By that measure, every company in the world except the magnificent seven is a niche curiosity.
664.
by senko 2y ago
> it's just a complete fantasy from the economist It's par for the course for Charlemagne (the column in The Economist). It's not meant to be an actual realistic analysis and reporting (the way their non-columnist articles
665.
by senko 2y ago
There's no single "style of PowerPoint". People use it for various, different, things, and if you look at a deck made for one purpose through the lens of another, it'll look atrociously bad. You can use PowerPoint (or Go
666.
by senko 2y ago
Notion is basically a wiki too, and seems pretty popular in startups. If you squint a bit, Obsidian as well.
667.
by senko 2y ago
> And no I've spent far less time "fixing" Windows. [...] I just had to reinstall the system from scratch. That's one way of dealing with it. > Never mind how once I had to fix a glaring issue in the Linux kernel w
668.
by senko 2y ago
> I'm a tad more humble. Mentioned more than a dozen times in the kernel git log. Exceptionally humble, indeed. > You've never said how much time you've wasted fixing Linux or making software work. This comes up often a
669.
by senko 2y ago
The US can't even make Puerto Rico a US state.
670.
by senko 2y ago
Is there a book covering TSMC in a similar way? (I know about Chip War but hesitant to spend time with it as I'm really not interested in politics of it all).
671.
by senko 2y ago
Like shelve, but with json instead of pickle, and sqlite instead of dbm; and you have a set of dicts instead of k/v store , with helpers to search by any dict field; and also pydantic support.
672.
by senko 2y ago
I've built this (unironically) for Python: https://github.com/senko/dante It combines the convenience of SQLite (nothing to install) with the convenience of just throwing stuff in a dict. Perfect for quick prototy
673.
by senko 2y ago
Since Disney classics often riffed off public domain works, the school conceivably might have gone straight to the source.
674.
by senko 2y ago
Linux kernel has been in development since the nineties, not just for the last ten years. Also 5000 contributors is a lot more than 2000 from gp's comment. Let's ignore the years before dotcom boom since the dev community was prob
675.
by senko 2y ago
> The two numbers mentioned in the article are 176 mil and 612 mil. Those two numbers are from the intro. The postscript and the updates at the end mention $1.4b and $3b respectively. The real cost is probably impossible to calculate, bu
676.
by senko 2y ago
The cost of Linux kernel development alone has been estimated at a few $B ( https://dwheeler.com/essays/linux-kernel-cost.html ), current figure is probably over tens of billions. Also, the prize doesn't require you
677.
by senko 2y ago
And Linux kernel, curl, SQLite and many other open source software are worth infinitely more than the purchase price. Also, you cut off the "from the benchmark" part; this doesn't expect it to solve any random Github issue, j
678.
by senko 2y ago
That's project management, not product management.
679.
by senko 2y ago
This describes the need to iterate on the product idea, in collaboration with engineering, instead of slinging vague PRDs over the wall. Table stakes for any product manager, not just AI related.
680.
by senko 2y ago
The original is, well, OG, but I prefer the Orbital version, esp. when listening on something that won't butcher the bass (ie not a phone, laptop, or cheap headphones).
681.
by senko 2y ago
There used to be more brands with trackpoints but IMHO most were pretty bad, except for Lenovo. That said, their newer ones also seem a bit more slippery / harder / less precise than older ones, in my (subjective) impression.
682.
by senko 2y ago
> a part-time remote job coupled with full time traveling on a budget a.k.a. Digital Nomad
683.
by senko 2y ago
This only contains V4 UUIDs. Disappointed.
684.
by senko 2y ago
For some reason you can't just add a custom search engine url to Firefox, you need to have an extension. Kagi has one that swts up the engine and injects the token, but it sometimes breaks so I need to uninstall/install it. Really
685.
by senko 2y ago
I do almost all browsing (except for services where I need to be logged in) from private mode. For the ones where the app needs to know me, I have separate profiles (multi-account containers ftw!). Nothing especially sensitive, I just consi
686.
by senko 2y ago
I'm happy with the price but Firefox randomly logs me out in private mode (maybe after an update? too infrequent to be able to spot the pattern, but it's there), which is a bit annoying.
687.
by senko 2y ago
They list Claude, OpenAI GPT, Mistral, Gemini and Llama: https://help.kagi.com/kagi/ai/assistant.html#llms-available-... It's a fair assesment, unless you're nitpicking about "all" in there.
688.
by senko 2y ago
> It's never going to generate anything new. It literally can't. It can't on it's own . But why does it need to? As a tool, the user can provide insight, imagination, soul, or guidance. And let's be honest, very
689.
by senko 2y ago
> I asked that to prove convolution of two gaussian is gaussian. [The solution] is correct but not elegant. The goalposts are moving at the speed of light. A few years ago if someone told us that you could ask a computer to compose a poe
690.
by senko 2y ago
Great article, Addy gets to the core of it and explains in a non-biased (pro or con), non-hype way. The examples, patterns and recommendations match what I've seen pretty well. I've been working on an agentic full-app codegen AI s

More ›