Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
theshrike79
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
91.
▲
by
theshrike79
2mo ago
As long as the only copy of the blog doesn't exist on Github, it really doesn't matter IMO.
92.
▲
by
theshrike79
2mo ago
Defuddle ( https://github.com/kepano/defuddle ) by @kepano (CEO of Obsidian) is pretty damn good at getting a text-only version of random sites.
93.
▲
by
theshrike79
2mo ago
It's 110% vibe engineered, so I'm rather hesitant to share it to anyone else - I don't need the drama =) But it's pretty much a bespoke handler for every site with some common bits (caching, feed fetching etc) For exampl
94.
▲
by
theshrike79
2mo ago
I built a RSS feed hydrator just for this. It grabs the basic feed and adds whatever I need to it. Then it generates a new one that I subscribe to on my feed reader.
95.
▲
by
theshrike79
2mo ago
There's no continuity with outsourced consultants either. But still people manage to make it work. With the exact same tools you use with LLMs: proper documentation and detailed instructions. Then you can grab any random consultant or
96.
▲
by
theshrike79
2mo ago
Soft skills isn't just about being nice. It's about knowing the capabilities of people, what can they do, what do they excel at and where they need support. The exact same applies to LLMs, each of them have their own very distinct
97.
▲
by
theshrike79
2mo ago
AFAIK it's going to be a public API from Anthropic. Dunno about the costs.
98.
▲
by
theshrike79
2mo ago
As Anthropic explains it, the watermark is at the model level. It just produces stuff that matches the watermark, prompts and instructions don't matter.
99.
▲
by
theshrike79
2mo ago
This is just managing the low-effort low hanging fruit. The students who, without a single neuron activating, will just copy paste questions to an AI web UI and paste the answers back. Or the mid-tier corporate people who file 50 page "
100.
▲
by
theshrike79
2mo ago
I’ve been running with Caveman mode since it came out and I haven’t seen any of this.
101.
▲
by
theshrike79
2mo ago
This always reminds me of the Trace Buster Buster Buster -scene from The Big Hit :D https://www.youtube.com/watch?v=Iw3G80bplTg
102.
▲
by
theshrike79
2mo ago
There are stories of students literally copy-pasting LLM answers to school assignments with prompts and all. This will most likely bring a sift end to at least the low hanging fruit.
103.
▲
by
theshrike79
2mo ago
Doing a good "deepfake" Photoshop requires skill. With an AI model it requires the ability to speak or write, not much more.
104.
▲
by
theshrike79
2mo ago
I have a friend who works in B2B eCommerce specifically with some projects in D2C and they're saying that SEO is mostly a dead end now. It's all about Answer Engine Optimization (AEO) or Generative Engine Optimization (GEO). So if
105.
▲
by
theshrike79
2mo ago
The monopoly movie chain in my country "updated" their pages to be "better". Now it's so bad that some chad created their own overlay for the site where you can actually see all the showtimes for a specific movie on
106.
▲
by
theshrike79
2mo ago
Read only? Yes. But the readonly needs to be enforced on the service side. Like my personal agent has read only access to my Fastmail account via their MCP. It can't send mail as me, but it can read, categorise and organise my mail. If
107.
▲
by
theshrike79
2mo ago
Counter point: Companies have made their pages as hard as possible to access with automated tools. People move to agents, if your site is "invisible" to agentic users (no API, no MCP) people won't find it. Companies start ope
108.
▲
by
theshrike79
2mo ago
I use mise for dev tooling, that way I can have the exact correct version for every project. Python through brew is the one I expect to be the latest one I use for one-off scripts.
109.
▲
by
theshrike79
2mo ago
> The auto mode classifier uses a small number of extra tokens per tool call, and we're no longer charging Claude Code users on Pro, Max, and Team plans for that classifier overhead, effective today. Not a very good business giving
110.
▲
by
theshrike79
2mo ago
This was something the Nokia N900 was excellent at. It was basically a Linux computer anyway so making it a server wasn't a massive undertaking.
111.
▲
by
theshrike79
2mo ago
What's the difference between "trash" and "tropes" in your mind? For example the good old Harlequin books were pretty much trash: https://en.wikipedia.org/wiki/Harlequin_Enterprises But when th
112.
▲
by
theshrike79
2mo ago
I think many people forget that the vast majority of books aren't high art. Your mind goes to Heinlein, Arthur C. Clarke, N.K. Jemsin, or Pratchett. My mind goes to Ice Planet Barbarians[0] or Crimson Worlds[1] (both written by actual
113.
▲
by
theshrike79
2mo ago
First to market usually wins. After they've got their gizmo in every user's hands, who will buy a perfectly crafted replacement with E2EE and all the bells and whistles? Shove it out as fast as possible, patch it live.
114.
▲
by
theshrike79
2mo ago
For me it's a dopamine machine. I get my doses of lacking dopamine from completing things. Agents let me complete things in parallel all the time. They also let me build tools that manage all the attention-grabbing things in my life. S
115.
▲
by
theshrike79
2mo ago
There were some changes when "american_patriot_47" style political accounts were discovered to be from Sri Lanka or India.
116.
▲
by
theshrike79
2mo ago
Jameela Jamil put this out pretty well: https://www.instagram.com/reel/DOvh4_UDHGy/ (sorry for the instagram link, couldn't find the original clip) Basically the gist of it is that if you want to be "on
117.
▲
by
theshrike79
3mo ago
Dunno if it's a generational thing or what, but I check my email like maybe once a day sometimes I can go 2-3 days without looking at it. It's just a place where Stuff gathers and I go through it when I Get around to it. Can'
118.
▲
by
theshrike79
3mo ago
My ADHD gets dopamine from getting something working, not from the tedious process of getting there. So YMMV.
119.
▲
by
theshrike79
3mo ago
Where and who is the pilot who crashed btw?
120.
▲
by
theshrike79
3mo ago
Subagents are good when the harness (And agent?) understand how smart they are. My root level CLAUDE.md has pretty much just "use a lower tier agent when relevant". Then I daily-drive Opus, it automatically offloads simpler stuff
More ›