coffeecoders

8 ms·

31.
by coffeecoders 10mo ago
I feel like being in a time loop. Every time a big company releases a model, we debate the definition of open source instead of asking what actually matters. Apple clearly wants the upside of academic credibility without giving away commerc
32.
by coffeecoders 10mo ago
My prediction: 2026 looks normal. AI stays the top story but in a boring way as novelty wears off and models get cheaper and faster (maybe even more embedded). No AGI moment. LLMs start feeling like databases or cloud compute. No SpaceX or
33.
by coffeecoders 10mo ago
I've realized over time that I personally cannot learn from video at all. Even "great" lectures don't stick. Text does! Being able to skim, jump around, re-read a paragraph or pause on a single sentence is how understand
34.
by coffeecoders 10mo ago
Charging by minute might push people toward shorter, noisier and more fragmented pipelines. It feels more like a lever to discourage selfhosting over time. It's not outrageous money today, but it's a clear signal about where they
35.
by coffeecoders 10mo ago
Us software engineers assume value comes from serving more people, faster, with less friction. But many of the things that actually make life feel coherent such as learning a craft, maintaining friendships and building tools for one person,
36.
by coffeecoders 10mo ago
I have an Apple TV and I’ve been running iSponsorBlockTV [1] on my Synology box for a while. It auto-skips the sponsored segments and with Youtube premium, it gives me a clean, ad-free setup. I can’t stand those in-video intros or sponsored
37.
by coffeecoders 11mo ago
This is good. It’s fascinating how it spins up interactive pages instantly. Some of the mini-apps actually feel useful, but others break in ways you wouldn’t expect. I’m curious to see how it evolves with more complex, multi-step queries.
38.
by coffeecoders 11mo ago
Not quite. This isn’t saying React or Next.js are fundamentally insecure in general. The problem is this specific "call whatever server code the client asks" pattern. Traditional APIs with defined endpoints don’t have that issue.
39.
by coffeecoders 11mo ago
This vulnerability is basically the worst-case version of what people have been warning about since RSC/server actions were introduced. The server was deserializing untrusted input from the client directly into module+export name looku
40.
by coffeecoders 11mo ago
Location: Vancouver, BC, Canada Remote: Yes (preferred) Willing to relocate: No Role (Current): Senior Full-Stack Engineer, Bay area. Resume/CV: Available on request Email: ntemail (at) protonmail.com Technologies & Expertise: - 10
41.
by coffeecoders 11mo ago
I am of a belief that upcoming winter will look more like normalization than collapse. The reason is hype deflation and technical stagnation don't have to arrive together. Once people stop promising AGI by Christmas and clamp down on i
42.
by coffeecoders 11mo ago
I don't know who said it, but it fits the current Windows vs Linux moment perfectly. "Most people don’t want control. They want optionally having control." Windows gives you the worst of both worlds. It restricts the control
43.
by coffeecoders 11mo ago
Funny thing: this feels "realistic" because it’s not perfect physics. A perfectly simulated Hooke's law spring actually looks fake and too stiff. But if you let the animation wobble a bit more and slow down the damping, our b
44.
by coffeecoders 11mo ago
No, visiting a URL does not automatically expose your exact latitude and longitude. I just get approximate location from your public IP address via an external IP geolocation API (ipapi.co), which usually gives city-level accuracy.
45.
by coffeecoders 11mo ago
Thanks! The Unique Fingerprint ID is basically a hash of all the collected fingerprint fields. [1] The Canvas Deep Fingerprint Hash is higher entropy and includes baseline shapes, emoji rendering, winding rules etc. [2]. It’s meant to captu
46.
by coffeecoders 11mo ago
Yeah, I made the mistake of including all features, even ones that change on every refresh like canvas or audio. The fingerprint should really only use stable features that don’t fluctuate between reloads. That way it’s consistent for the s
47.
by coffeecoders 11mo ago
You're 100% right. The raw fingerprint dumps alone are not actually useful unless you can compare them to a population. And creating that comparison is far harder than people think. To answer "How unique am I?" I need a large
48.
by coffeecoders 11mo ago
If we all rely on others to skim /newest, the whole curation system collapses. Maybe the homepage should surface a couple random fresh posts too?
49.
by coffeecoders 11mo ago
Interesting approach, but I mean more in the sense of a multi-agent sandbox than workflow automation. Your project feels like wrapping a bunch of LLMs into "agents" with fixed cadences, it is a neat product idea, even if it mostly
50.
by coffeecoders 11mo ago
How hard would it be to run a simulator with multiple LLMs. Say, one as the boss and a few as employees. Just let them talk, coordinate, and "work"? Could be the fastest way to test what actually happens when you try to automate m
51.
by coffeecoders 11mo ago
Yes. That particular content-farm business model (rewrite 10 articles -> add SEO slop -> profit) is effectively dead now that the marginal cost is zero. I’m not mourning it.
52.
by coffeecoders 11mo ago
I’m not sure I agree with this sentiment. You can type something "by hand" and still have almost no voice in it if the incentives push you to flatten it out. A lot of us spent years optimizing for clarity, SEO, professionalism etc
53.
by coffeecoders 11mo ago
I actually think we’re overestimating how much of "losing our voice" is caused by LLMs. Even before LLMs, we were doing the same tweet-sized takes, the same medium-style blog posts and the same corporate tone. Ironically, LLMs mig
54.
by coffeecoders 11mo ago
I think the real story isn’t that AI will replace 11.7% of workers. It is that we are about to discover that far more than 11.7% of the work we do was never actually work in the first place. Workflows that were untouchable will now be overh
55.
by coffeecoders 11mo ago
Hi HN, I’ve been experimenting with ways to reduce my browser fingerprint and exploring techniques to anonymize fingerprint data. So I built this. This is kind of like a lighter, more thorough version of CreepJS but entirely client side. I
56.
Show HN: Explore what the browser exposes about you(neberej.github.io)
275 points by coffeecoders 11mo ago|90 comments
57.
by coffeecoders 11mo ago
Feels like the same consolidation cycle we saw with mobile apps and browsers are playing out here. The winners aren’t necessarily those with the best models, but those who already control the surface where people live their digital lives. G
58.
by coffeecoders 11mo ago
I bought an Ipad yesterday. The setup was almost the same, except all the bloat was from Apple itself. Numbers, GarageBand, iMovie, Keynote, Pages, Clips… even the Tips app felt like bloatware. Wild how every ecosystem has its own "pre
59.
by coffeecoders 11mo ago
LLMs are terrible for out-of-distribution (OOD) tasks. You should use chain of thought suppression and give constaints explictly. My prompt to Grok: --- Follow these rules exactly: - There are 13 hours, labeled 1–13. - There are 13 ticks. -
60.
by coffeecoders 11mo ago
Napkin math: 9 AIs × 43,200 minutes = 388,800 requests/month 388,800 requests × 200 tokens = 77,760,000 tokens/month ≈ 78M tokens Cost varies from 10 cents to $1 per 1M tokens. Using the mid-price, the cost is around $50/mont

More ›