HackerTrans
TopNewTrendsCommentsPastAskShowJobs

talolard

no profile record

Submissions

Agent Skill to Use a Debugger

github.com
2 points·by talolard·4 miesiące temu·2 comments

The AI-Powered 10-Minute Habit That Taught My Kid to Read

talperry.com
3 points·by talolard·5 miesięcy temu·2 comments

Five Practical Lessons for Serving Models with Triton Inference Server

talperry.com
23 points·by talolard·6 miesięcy temu·6 comments

comments

talolard
·4 miesiące temu·discuss
I’ve been vibe coding so much I didn’t even remember how useful this could be
talolard
·5 miesięcy temu·discuss
I used DallE and Anki to teach my son to read. I'm not sure the HN crowd will be wowed by old tech, but this post leaves techs to the agents, and deals with the emotional aspects of learning I learned.

Yes, I wrote the clickbaity title with AI, but the rest of the post by hand.
talolard
·6 miesięcy temu·discuss
I haven’t had major issues with sccs yet. The linter enforces forward references so the cycle pain we do have is with dynamic/deffered imports, and it’s usually solved by splitting a module.

If you look at the pyrefly repo (metas new type checker), there are some deep thoughts about sccs, but I didn’t fully grok them.
talolard
·6 miesięcy temu·discuss
Not IAC, but I’ve been doing a similar trick to sequence adding type annotations to python code,

Eg take the module graph, break the SCCs in a similar manner , then take a reverese topological sort of the imports (now a dag by construction).
talolard
·6 miesięcy temu·discuss
Hi HN,

I’m wrapping up a role where I spent a significant amount of time writing Triton kernels. It’s a fantastic tool, but the learning curve has some sharp edges. I wanted to share a few practical "notes from the field" for anyone moving beyond the very opaque docs.
talolard
·5 lat temu·discuss
I have a few suggestions regarding search performance.

As others have mentioned, there isn't one "right answer" but there are a few nifty tricks you can use to boost performance.

You could try a trigram index[0] as the primary search index. You'd loose stemming, but gain language support and indexed regular expression search which is a nice power user feature.

Looking at your where clause, did you index language and textsearchable_index_col together ? If not, you are giving the planner the opportunity to skip your full text index.

if you order by ts_rank_cd then the query will sort all rows, including those that don't match. A better pattern is to take the ranking score as a column and then sort it in a subquery. [1] from stack overlfow has an example. (As an aside, from pg2, CTEs are no longer an optimization fence, so you can write the query out more cleanly with a CTE and still get the desired performance).

You should experiment with GIN vs GIST indices. GIN are larger, so on a large dataset the index might not fit in memory. You could pay more to have more memory, but worth trying a GIST index to see if that makes things faster just because it fits in memory.

A final frontend comment, I'm a fan of infinite scroll for this kind of stuff. You already have pagination effectively set up, you could spoil me as a user with infinite scroll. react-virtualized[2] is a great library for that.

Hope that helps.

[0] https://www.postgresql.org/docs/9.6/pgtrgm.html [1] https://dba.stackexchange.com/questions/4054/ts-rank-massive... [2] https://github.com/bvaughn/react-virtualized
talolard
·6 lat temu·discuss
This is great! We have a toddler and it’s great to know something like this exists.

As a suggestion, I’d rather buy a 5 or 10 pack rather than a monthly subscription. We don’t need this every month but expect to use it more than once