kulikalov

8 ms·

1.
by kulikalov 5mo ago
Md and plotly is all you need. The only thing that is truly missing is some sort of Markdown-based forms
2.
by kulikalov 5mo ago
I think this is exciting. The market will do its job and crush the inefficient companies where management is unable to recognize the slop. People who produce value will produce more of it with AI, people who wasted resources will waste more
3.
by kulikalov 5mo ago
Slop architecture leads to compounding problems that people try to solve with more slop. If one wants to control the quality of the code then the throughput and multithreading is bottlenecked by how much code one can comprehend in a given p
4.
macOS Notifications for Claude Code and AeroSpace(kulikalov.com)
1 points by kulikalov 6mo ago|0 comments
5.
by kulikalov 6mo ago
Or it could be a selection bias. The ground truth is not what HN herd mentality complains about, but the usage stats.
6.
by kulikalov 6mo ago
I love how easy this kind of mini data visualization projects have become
7.
by kulikalov 7mo ago
“Social trust in autocracies”
8.
by kulikalov 7mo ago
I came to the exact same pattern, with one extra heuristic at the end: spin up a new claude instance after the implementation is complete and ask it to find discrepancies between the plan and the implementation.
9.
by kulikalov 9mo ago
Economic Facts and Fallacies by Thomas Sowell
10.
by kulikalov 9mo ago
Loved it too! Try echopraxia next.
11.
by kulikalov 10mo ago
The right way to shape mass preferences is to collectively decide what's right and then force everyone to follow the majority decision under the muzzle of a gun. <sarcasm off> Did I capture the sentiment of the hacker new crowd
12.
by kulikalov 11mo ago
AI Copilot for brick-and-mortar retailers ( https://fastquery.ai ) Perfect usage metrics. Store staff spends 35hr per week using my software in a ~7 employee per shift setting. No churn. Bootstrapped past $100k ARR on my own, just
13.
Ask HN: How to remove space switch delay on macOS?
3 points by kulikalov 1y ago|1 comments
14.
by kulikalov 2y ago
russia hasn't been cutoff economically still. 3 years and the economy is going down but not as fast as it should have. The leaders of the west think they are playing a game here with pros and cons. There are no pros and a huge con with
15.
by kulikalov 2y ago
Sounds inefficient. It’s like predicting the boiling point of a kettle by measuring the speed of individual molecules of water.
16.
A multimodal dataset with one trillion tokens(github.com)
224 points by kulikalov 2y ago|52 comments
17.
by kulikalov 2y ago
the question is - why lifts require windows?
18.
by kulikalov 2y ago
Looks great! I'd love to get a notification once you add remote servers support. What channel should I use? Besides discord.
19.
by kulikalov 3y ago
I like the article and am a big proponent of simple architectures. But. The described use cases have homogenous workloads. Hence simple architecture. Extrapolate the rest. And of course there are for sure cases of people increasing complexi
20.
by kulikalov 3y ago
It happened 15 years ago. Half of the town knew that, all the police knew that. In fact the police tortures people themselves, it's not news for them. This thread is in the midst of discussion of a public person tortured for 3 years an
21.
by kulikalov 3y ago
There's no cutoff. Only some inconveniences for population maybe.
22.
by kulikalov 3y ago
Parts of population always did and always will be falling behind other parts. Humanity must ensure that an individual has the way to realize their potential. Freedom to raise and freedom to fall. The internet changed everything. The informa
23.
by kulikalov 3y ago
There are smart people who work for Russia, or worse - for russian gov. Without them it would not be possible. How can an individual have both the critical thinking and still have the gut to contribute to this? I grew up in rural russia. It
24.
by kulikalov 3y ago
Why? Because there's no word for "insurance of AI advise accuracy"? The whole point of progress is that we create something that is not a thing at the moment.
25.
by kulikalov 3y ago
Well, I guess since it's not "practice" we gonna call it "mal-inference insurance".
26.
by kulikalov 3y ago
ChatGPT enterprise? Or over API. They state that those offerings data is not used for training. I'm not a lawyer but afaik illegally retrieved evidence cannot be used - "exclusionary rule".
27.
by kulikalov 3y ago
It's funny how any conversation ends with this question unanswered.
28.
by kulikalov 3y ago
Given orders of magnitude better cost efficiency, they will have plenty of funds to lure in any insurance firm in existence. And then replace insurance firms too.
29.
by kulikalov 3y ago
Can't a tech firm running a "legal gpt" have an insurance?
30.
by kulikalov 3y ago
Andrej Karpathy videos on YouTube and Stanford NLP book.

More ›