Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
sakesun
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
121.
▲
by
sakesun
1y ago
Let's call it Premix
122.
▲
by
sakesun
1y ago
Here is the first announce of doctest I can find (1999) I think it's Python 1.5 time. https://groups.google.com/g/comp.lang.python/c/DfzH5Nrt05E/m...
123.
▲
by
sakesun
1y ago
Lot of downvote at my message. You are probably the only person who understand my joke :)
124.
▲
by
sakesun
1y ago
Letting AI control software releases might seem risky, but it may not be as problematic as it sounds—especially if you also use AI to handle issue support. :)
125.
▲
by
sakesun
1y ago
Wow
126.
▲
by
sakesun
1y ago
We need better tool to understand the system. But, before that, please spend sometime to understand the different between gtoolkit and lepiter.
127.
▲
by
sakesun
1y ago
Similar to how some domain name sellers acquire desirable domains to resell at a higher price, agent providers might exploit your success by hijacking your project once it gains attraction.
128.
▲
by
sakesun
1y ago
Similar to Logseq. They are both built with Clojure.
129.
▲
by
sakesun
1y ago
Something to do while AI take care of other chores.
130.
▲
by
sakesun
1y ago
Credit should also go to Rye project by Armin Ronacher.
131.
▲
by
sakesun
1y ago
Hinted by this article, next version of Gmail system prompt might craft system prompt specifically for the author, with insight even the author himself not aware of. "You're Greg, a 45 year old husband, father, lawyer, burn-out, n
132.
▲
by
sakesun
1y ago
It's more complex that decade ago, but still a relatively simple language. I can understand the article without much effort, while I scratch my head really hard when read about advance feature of Typescript or Scala.
133.
▲
by
sakesun
1y ago
Soon, some free smart LLM code generators will stop generating certain outputs and instead suggest using commercial components that have paid for promotion.
134.
▲
by
sakesun
2y ago
I heard similar things about search engine back in 2012
135.
▲
by
sakesun
2y ago
A digital rishi.
136.
▲
by
sakesun
2y ago
When human do not pass AI inverse-Turing test.
137.
▲
by
sakesun
2y ago
I always think the coming of this feature is inevitable.
138.
▲
by
sakesun
2y ago
> Knows Perl/lisp/zig/ruby/C++/python, in order of most to least proficiency Whoa. A highschooler who prefer Perl & Lisp.
139.
▲
by
sakesun
2y ago
Even worse there are groups of people who keep praising it and keep us curious through these years. Yet none of remarkable applications built with it except the tool itself.
140.
▲
by
sakesun
2y ago
Project motto: "Make systems explainable through <unexplainable>contextual micro tools</unexplainable>"
141.
▲
by
sakesun
2y ago
While I think what proposed will never happen, this item is full of interesting debate.
142.
▲
by
sakesun
2y ago
I doubt their vision. If it's for personal coding project, who would pick RSC ?
143.
▲
by
sakesun
2y ago
Never know we can format comment like this in HN :o
144.
▲
by
sakesun
2y ago
> cuOpt will be released as open source and freely available for developers, researchers and enterprises later this year.
145.
▲
by
sakesun
2y ago
Arroyo is another one based on DataFusion
146.
▲
by
sakesun
2y ago
Next version of MS Recall/Copilot will watch the code you type in any IDE or editor and try to give compiler warning automatically.
147.
▲
by
sakesun
2y ago
I have been told and believed that we are getting very close to the ultimate answer throughout my career since CORBA/DCOM. Just learn that there is another discontinued attempt https://serviceweaver.dev/
148.
▲
by
sakesun
2y ago
Your comment inspire me to brush up my Delphi skill.
149.
▲
by
sakesun
2y ago
I'm still amazed by how intelligent the outcome is, after these number crunching processes. Really cannot relate its ability to generalize information to the theory behind it.
150.
▲
by
sakesun
2y ago
I've found none of the explanations of how LLMs are built have been satisfying, especially considering how impressive the applications of them are.
More ›