senko

17 ms·

301.
by senko 9mo ago
Ignoring the " A high-level athlete doesn’t need to see a sports doctor " faux-pas, the article's biggest sin is equating "running" with long distance (marathon or ultra). To recreational runners, the more interesti
302.
by senko 9mo ago
Responses API is a commodity. That's why OpenAI tries to push Assistants API, Agents SDK and ChatGPT Apps which are more of a lock in: https://senkorasic.com/articles/openai-product-strategy-2025 Funny thing is, e
303.
by senko 9mo ago
As a heavy Claude Code user, I would like to have that option. But if it's just 33% as good, I wouldn't bother. Top LLMs have passed a usability threshold in the past few months. I haven't had the feeling the open models (fro
304.
by senko 9mo ago
I try to balance being informed and ignoring the mostly irrelevant hot topics du jour. My approach: * world: weekly Economist coverage of world/biz/general topics (audio via app) - keeps me generally up to date * local: daily dige
305.
by senko 9mo ago
Sure, that works as well, for example on some deploys I set the settings in systemd service file. However, it's more convenient to just have .env right there. > On production keep things like API keys that need to be kept secret els
306.
by senko 9mo ago
Possibly. If the installer happy path fails (which has happened to me), Arch is "here's root shell, figure it out", Ubuntu is slightly more user-friendly :) I will say Arch wiki is amazing, even if you're not using it. I
307.
by senko 9mo ago
Yeah agree completely. Started with Slack at around 2.x, took a long time to switch to Debian and occasionally Ubuntu. That was when you compiled your own kernel and installed software by running ./configure && make &&
308.
by senko 9mo ago
Arch is great. However, I would never recommend Arch (or an Arch derivative) to a first-comer to Linux. Ease in gently, with Ubuntu or Fedora. Get familiar. Then go crazy.
309.
by senko 9mo ago
On Linux, there are 78 dynamically linked libraries, such as for X11, vector graphics, glib/gobjectlibgobject, graphics formats, crypto, encryption, etc.
310.
by senko 9mo ago
You need to know "qrencode" exists under that exact name. Claude already knows about it and how to use it.
311.
by senko 9mo ago
Some more Django recommendations from Frank Wiles of Revsys (Django consultancy): https://frankwiles.com/questions/starting-django-projects/ I'll add a few of my own: * Set up the project using uv * I name th
312.
by senko 9mo ago
You still need clear separation between frontend and backend (react server components notwithstanding), so nothing's stopping you from using Python on the backend if you prefer it. Django with DRF or django-ninja works really nice for
313.
by senko 9mo ago
You can do three stage: 1. Make a schema migration that will work both with old and new code 2. Make a code change 3. Clean up schema migration Example: deleting a field: 1. Schema migration to make the column optional 2. Remove the field i
314.
by senko 9mo ago
> The end goal of AI is AI is a technology. It has no goal. You use a tool, the tool doesn't use you or have goals or plans for you. > In a world where there is no work for the common man. "Work expands to fill the time avai
315.
by senko 9mo ago
You're right, they shouldn't have started a company, that would be better for diversity.
316.
by senko 9mo ago
> There must a ton of new full-web datasets out there, right? Sadly, no. There's CommonCrawl ( https://commoncrawl.org/ ) which still, sadly, far removed from "full-web dataset." So everyone runs their own s
317.
by senko 9mo ago
On a tangent: the origin of the problems with low-quality drive-by requests is github's social nature. That might have been great when GitHub started, but nowadays many use it as portfolio padding and/or social proof. "This p
318.
by senko 9mo ago
> The dev dependency group is a special case for uv run - it will always install those dependencies as well Huh, this is...unexpected. TIL, indeed!
319.
by senko 9mo ago
I don't mean to say it's trivial. I'm sure there are many hard problems such as the one you mention - though that particular one is more "cleaning the index" part which might work on top of the open common corpus. B
320.
by senko 9mo ago
Yes. But they don't crawl everything (probably due to lack of funding), and , as the article and other commenters here note, people are incentivised to allow Google and only Google to crawl. In practice, the CommonCrawl dataset is t
321.
by senko 9mo ago
A full up-to-date index of the searchable web should be a public commons good. This would not only allow better competition in search, but fix the "AI scrapers" problem: No need to scrape if the data has already been scraped. Craw
322.
by senko 9mo ago
> Your robots.txt file is the very first thing Googlebot looks for. If it can not reach this file, it will stop and won't crawl the rest of your site. Meaning your pages will remain invisible (on Google). This implication (stopped
323.
by senko 9mo ago
> It was morbidly fascinating watching it rabbit hole. Oh yes, it's like slow-motion train wreck. Once it gets off the rails, the best thing is to nuke the context, reset the state (git reset --hard or equivalent) and retry, this ti
324.
by senko 9mo ago
> It’s not that the agents are now producing flawless code. I spent a good 20 minutes yesterday watching one tie itself in knots trying to write a regex: first in Sed, then in Bash, and finally in Python (six times). This sounds very s
325.
by senko 9mo ago
Had a few Xiaomis and a few Pixels (9a now). Pixels are better, but more expensive. Xiaomi bangs/buck was pretty good. Lately they've been upping the price so nowadays they're not clearly the budget option, but a few years ag
326.
by senko 9mo ago
> it was Nokia (Pre-Elop) Nokia fan here. The S40 devices were awesome, sturdy, trusty tools. Smartphones today are light-years more capable, but the constant churn, treating users like idiots, and the incidental complexity of modern pho
327.
by senko 9mo ago
Ahh, a topic dear to my heart. > Thinkpads are repairable [...], durable [...]. What's not to like? > Thinkpads are “Cheap” because nobody who worships the things buys one new. Confession time: over the years, I've bought 4:
328.
by senko 9mo ago
Chrome can run small[0] models and can auto-download them: https://developer.chrome.com/docs/ai/built-in [0] "small" in comparison to ChatGPT, but still a bulky download
329.
by senko 10mo ago
Not sure where to report this, so here goes: something weird's going on with the title setters (from default .bashrc, the 'case' block near the top) when connecting from my Linux box. Tried with xterm, tilix and ghostty, all
330.
by senko 10mo ago
Let's rephrase this a bit so I'm sure I understand. Microsoft , a $3.4T company, should charge people for open source they didn't even write? Hell no. Hell no .

More ›