bel8

4 ms·

1.
by bel8 5d ago
Either dark pattern or incompetence. Sometimes both.
2.
by bel8 5d ago
For me it works at first but Claude models forgets it after some prompts, despite only using like 100k tokens.
3.
by bel8 7d ago
In the context of PHP it is notable to be able to run a production grade web server without nginx/apache/docker in front of it. PHP doesn't have a production grade built-in server like .NET's Kestrel.
4.
by bel8 7d ago
OS/Software is why I don't like macs. For me it's like they built a Ferrari with 4 thin bicycle tires. What a waste.
5.
by bel8 8d ago
They probably have an explicit instruction in their own CLAUDE.md to not implement that support.
6.
by bel8 8d ago
Kinda. Not my engine but this entire engine fits inside the context window with tons of tokens to spare: https://github.com/KilledByAPixel/LittleJS Videos of example games: https://www.youtube.com/@Kill
7.
by bel8 9d ago
to make it even more egregious: "Quick! Climb to about 1000 feet above sea level and the trex will boil and we will be safe!"
8.
by bel8 11d ago
interesting. So less specialization on one side but also less fragmentation.
9.
by bel8 11d ago
this is great! congratulations. one question I have is: since even rng has to be locally predictable won't things like lootboxes be hackable by reverse-engineering local executable? maybe for these cases the game could implement an API
10.
by bel8 12d ago
They don't have incentives to reclaim space imediately since disks that appear to be full sell more iCloud subscriptions.
11.
by bel8 12d ago
Imagine the pressure that an employer can impose onto H1Bs now, with no grace period.
12.
by bel8 12d ago
It's really powerful. I have been using it in Chrome for a while. This is how to enable: chrome://inspect/#remote-debugging
13.
by bel8 12d ago
> They didn’t invest in Siri... What makes you think that? They insisted on Siri for so long and even made a big deal of Apple Intelligence in WWDC. Then Apple threw the towel and waved the white flag with regards to Siri and replaced th
14.
by bel8 12d ago
Genius would have been for Apple to get Siri in a better state after so many years. This looks more like a participation medal, if at all.
15.
by bel8 12d ago
there's a setting to disable that
16.
by bel8 14d ago
I'd love to see these: - DeepSeek V4.1 Flash - Kimi K3 - GLM 5.3 (and flash) - hy4-preview - Grok 4.6 All of these can be acessed using a $10/mo OpenCode Go subscription.
17.
by bel8 14d ago
I'm not. Yes we normalized 1m context window and models tend to hallucinate less. But models have been somewhat stagnant since Opus 4.6/7. And in some regards there were even regressions like Claudeisms that are load bearing.
18.
by bel8 15d ago
Those were just examples. I never said it was the only thing happening. And asking to filter out the bubble you don't happen to like doesn't look like a solution.
19.
by bel8 15d ago
I like HN precisely because it tends to tell me what's happening. GitHub or AWS down? New LLM model? Some Apple event? Some famous tech person passed away? Some language achieved 1.0? I think removing what's currently hot opens a
20.
by bel8 16d ago
their screen are made by Samsung so the tech is known.
21.
by bel8 16d ago
dam I'll have to try that one
22.
by bel8 16d ago
It's an iPhone with a foldable screen made by Samsung. So it's hard to inovate here.
23.
by bel8 16d ago
It's not the worst status symbol. It's useful.
24.
by bel8 17d ago
in 2026? LLM performance, for one.
25.
by bel8 17d ago
Cross-platform is a thing. If anyting, Apple stands to gain from what was already developed for Android. Like https://developer.android.com/phones-tablets-foldables Which targets iOS also: https://kotlinlang.org&
26.
by bel8 18d ago
And it probably takes longer for a human to compile all that info.
27.
by bel8 18d ago
Poisoning LLM? I remember at least one study about it. Perhaps but it would have to be very subtle otherwise they might flag your content and shadowban it from future training. My personal opinion is that trying to make LLMs inefficient is
28.
by bel8 18d ago
I think they had this sitting idle for a while, waiting for their AI division to deliver a capable LLM to power it. Now they have Muse Spark 1.3, which is somewhere between Sonnet and Flable (Opus tier in my experience) in intelligence, but
29.
by bel8 18d ago
yup. GPT Luna tends to keep things objective. Muse Spark 1.3 is also one of the better models in this aspect, for me.
30.
by bel8 18d ago
you're right. ads are already infested with AI videos which means it works for some and those are no longer paying the price of fully manual editing.

More ›