yipinwong

7 ms·

61.
by yipinwong 1mo ago
Agreed. with "Ultra" (higher than Max), the luna performs really well for my non-metric-backed personal experience
62.
by yipinwong 1mo ago
The analysis is still not compelling for me to switch from gtp5.6-luna to GLM-5.3-flash given - costs per task $0.05 vs $0.09 - speed 130 vs 88 - where GLM has only 5 more intelligence point: at this point few point is meaningless for most
63.
by yipinwong 1mo ago
I'd still pay for your brain (expertise, experience for built them), which is separate from your agent :)
64.
by yipinwong 1mo ago
I don't know if that's the standard pricing for US models to go down overtime, while Chinese ones go up (start cheap but pay more). I don't have enough metrics to compare those costs but still Chinese models have been cheaper
65.
by yipinwong 1mo ago
Only those who mastered the craft makes their work look simple. The AI that wrote this might be the master not the writer, as this looks written by AIs. I will use the author's agents, not read his articles or use him for the job.
66.
by yipinwong 1mo ago
When reading this type of announcements, always have keen eyes on graphs. e.g. "Agent Coding Performance by Effort Level" cuts Y-axis from 0~20. - This makes it as if GLM-5.3-Flash made a bigger jump than it claimed as the Y-axis
67.
by yipinwong 1mo ago
Yes. AWS has OpenSearch for cloud-first vector searches (disregarding cloud-hosted postgres with extensions etc). It's ridiculously expensive and not as fast. DuckDB to the rescue. Also embed the crap outta DuckDB locally in local agen
68.
by yipinwong 1mo ago
Ox Alpha was working good for me but I do not it if a trend starts where openrouter hides where the traffic is going to.
69.
by yipinwong 1mo ago
Congrats to the DuckDB team. I've been eyeballing DuckDB and LanceDB as part of AI agent memories. This gives me a vibe that AWS will use DuckDB somehow in their ai agents sometime in the near future after seeing the potential.
70.
by yipinwong 2mo ago
ty for the concern. I'd like to back up my belief and claims based on my metrics and experience. If data backs it up, then I will apply my baynesian thinking to change my mind. Anecdotes normally triumphs in real life and biz, but not
71.
by yipinwong 2mo ago
I've seen so many I am not from China comment in all of the countries I mentioned like in HK, Japanese, Korea, Malaysia, but all turns out they are. Some of those countries communities expose countries, and yes, they are from mostly Gu
72.
by yipinwong 2mo ago
China follows foot-in-the-door tactic per the Art of War. (done in HK, Korea, Singapore, and Malaysia) This in-turn later on, AIs will train to make pro-China comments as AIs train on these. They got the sheer man-power, and with AIs it&#x2
73.
by yipinwong 2mo ago
The "wild" makes up for the 99% of China. Even in cities, why do I have to take a dump without doors? Why do I see mom's letting kids take a dump and pee in the streets whiloe they are wearing pants that have no coverage in t
74.
by yipinwong 2mo ago
Sounds like you live in top 1% area in China. That's def not true that toilets are readily available in China. As non-chinese citizen, I veered off place I cannot go without the passport, I see crappy areas without toilets everywhere
75.
by yipinwong 2mo ago
I do understand the sentiment of applying "constraints" over AI usage to certain scopes of our development work. This has an issue in practice. As an example, Function Programming (FP) has a high constraint on programming, where s
76.
by yipinwong 2mo ago
Real-life usage of interchangeable or synecdochical word triumphs in real life. My take on the post is for engineering disciple where JoeJag wants to create a common word while tackling "Agent" issues. I like Joe's approach a
77.
by yipinwong 2mo ago
You sound like a manager, not an engineer.
78.
by yipinwong 2mo ago
People ask XY problems and your job is to find what issue they are having, not try to help them with the attempted solution. Basically a StackOverflow guidance on XY problem applies to the general problem solving as well.
79.
by yipinwong 2mo ago
Boring is subjective. I like the article but was boring AF.
80.
by yipinwong 2mo ago
What I got was, be more generalist and become a polymath. My thoughts below --- This means, we need to read more to produce more in many different fields, and domains. The idea is also captured in today's top 10 article, "To becom
81.
by yipinwong 2mo ago
I think there are a few common themes to the failure reasons, but cannot get my hands on it. This seems like a list of reasons while I am looking for more abstract directions on how to prevent them. --- I am trying not to use AIs to just do
82.
by yipinwong 2mo ago
Good writing. Maybe I am too used to OREO format, where point is made, example given, and point re-iterated in tech or non-fictional format unlike the author's more of story-based narrative. The author seems to be branch off to a polym
83.
by yipinwong 2mo ago
The site layout is great for reading, so props to the author. I understand the sentiment that TUIs is not as accessible, sucks to use. I still do not like the fact that we should "stop making TUIs" altogether. Let me give you an
84.
by yipinwong 2mo ago
Explained simpler, there are crimes everywhere. Some tracked, some not. Untracked ones are the ones we track via Flock cams. There is only so much man-power that police or the govt can do.
85.
by yipinwong 2mo ago
Did I stutter? Have I mentioned anything China? WTF is that from? You are twisting my words with whataboutism. Stop digressing
86.
by yipinwong 2mo ago
Been years since "Ghost job" was to be regulated. Nothing's been done for years. This is yet another "they wanna ban" but won't do type of article. Sick of this.
87.
by yipinwong 2mo ago
Don't reply by just reading the title of the article. Read to the end. The author said "I thought retiring in my mid 30s after a few exits would be fun but". You never read the whole thing, thus don't get the context of
88.
by yipinwong 2mo ago
Check out some NIXOS communities. People they are doing fun crazy stuff, as NixOS is immutable, thus creates separate sandboxes (building/taking down) per agent, etc. I don't know most of the stuff there, but at least you can dive
89.
by yipinwong 2mo ago
This is a type of article that is HN worthy and came to HN for initially because I don't even understand a third of content there. Giving me inspiration to dive deeper. SEriously, I don't understand it (yet) lol.
90.
by yipinwong 2mo ago
The article does not relate AI usage for homework to exam score drops. It tries to link them together with no obvious cause and effect. The score could have dropped due to worse environment at schools or kids who were stuck at home studying

More ›