mmaia

Quick feedback: a lot of the job descriptions are Markdown that gets displayed as plain-text. If you rendered them it would be a lot more readable.

On the backend side, Cal Henderson published the details of Flickr's architecture in "Building Scalable Web Sites". It was a reference at the time, the "Designing Data-Intensive Applications" of another generation.

Interesting concept.

I've been doing the same kind of drill with ChatGPT. The disadvantage is the interface, but I can direct it to focus on what I'm learning or want to reinforce, etc.

> If your work isn't ready for users to try out, please don't do a Show HN. Once it's ready, come back and do it then. Don't post landing pages or fundraisers.

https://news.ycombinator.com/showhn.html

That's a cool tool.

FYI: There's a typo in the URL you posted, an extra `n` :)

Exactly. The post screams about all the issues I've seen in multiple companies between DS/MLE and SE/DevOps.

A suggestion for addition:

> Plausible Analytics is a standard Elixir/Phoenix application backed by a PostgreSQL database for general data and a Clickhouse database for stats.

https://github.com/plausible/analytics

A characteristic of the field since the beginning. Reading What Computers Can't Do in college (early 2000s) was an important contrast for me.

> A great misunderstanding accounts for public confusion about thinking machines, a misunderstanding perpetrated by the unrealistic claims researchers in AI have been making, claims that thinking machines are already here, or at any rate, just around the corner.

> Dreyfus' last paper detailed the ongoing history of the "first step fallacy", where AI researchers tend to wildly extrapolate initial success as promising, perhaps even guaranteeing, wild future successes.

https://en.wikipedia.org/wiki/Hubert_Dreyfus's_views_on_arti...

When asked to provide real life examples from the paper conclusion:

Workplace Example

Scenario: An employee has a colleague who tends to send aggressive emails if they don’t receive updates on time.

Healthy Active Avoidance: The employee learns, “If I send a quick status update every morning, I avoid the stress of hostile emails.” They adopt this as a habit.

Depressive Active Avoidance Deficit: A person with depressive symptoms may take longer to make this connection or fail to act even after realizing it. They know sending updates might help, but initiating the behavior feels too effortful or pointless. As a result, they keep receiving stressful emails, reinforcing the feeling of helplessness.

Published: October 8, 2014

More recent work indexed in Google Scholar: https://scholar.google.com/citations?hl=en&user=TVozLNoAAAAJ...

From the YC Startup Directory for W24 [0]:

USA: 195 UK: 6 Canada: 4 France: 4 Germany: 2 India: 2 Nigeria: 2 Brazil: 1 Kenya: 1 Sweden: 1

0 - https://www.ycombinator.com/companies?batch=W24

Starlink is available in Brazil and dominates the broadband market in Amazonas. Antennas are commonly found with illegal miners. [0]

0- https://www.bbc.com/portuguese/articles/cv2edkw84zmo (Portuguese)

https://www.starlink.com/map https://www.starlink.com/br/business https://www.starlink.com/br/residential https://www.starlink.com/br/service-plans

I tested it and liked the result. Good job.

A couple points:

  1. I couldn't find the option to generate just HTML with Tailwind (without React). Am I missing something?
  2. It took me some time to realize that you have a free plan. I think you should highlight that.

The best alternative I found is render.com

It has this feeling of "just works", and documentation is pretty good.

tldr

> While building a feature, we performed a database migration command locally, but it incorrectly pointed to the production environment instead, which dropped all tables in production.

Congrats on the achievement.

One trick part when working on "planet-scale" is parsing and matching the results for multiple countries. I tried some addresses in Brazil without success. Queries like "Starbucks Sao Paulo" return some results but addresses like "Avenida Paulista 100" (or its variations) don't.

Last time I looked (~2018) pelias-parser used some ML training and the results weren't very good for Brazil. I'm guessing in 2024, an open-source fine tuned LLM would do a good job?

Given that the architecture has a semantic layer, you just need to pick one that integrates with DuckDB, e.g. Cube [1].

About distributed DuckDB, have you checked Boiling Data? [2]

1- https://cube.dev/blog/introducing-duckdb-and-motherduck-inte... 2- https://boilingdata.medium.com/lightning-fast-aggregations-b...

Cool, I built something similar a while ago https://news.ycombinator.com/item?id=35918195

Happy to chat about it (twitter in profile).

> This doesn't really have a good market fit; people who don't care about these things just use google

Counterexample: Mapbox [1]

1- https://en.wikipedia.org/wiki/Mapbox

I like that it's not limited to single prompts and allows to have chat messages. It would be great if `OpenAIChatExperiment` could also handle OpenAI's function calling.

Looks like it's getting some fields wrong. For examples, "Global" location is matching other text not related to the job location.

Still, a good example of OpenAI's functions. I was using direct prompt for YAML parsing before and migrated to functions for my job board https://aijobnetwork.com/

I'm also using OpenAI's embeddings for recommendations: https://medium.com/@mauricio/how-to-add-recommendations-to-a...

Yes, that's actually my case. But, getting a salary on par with your skill set may require you to work directly with the employer or a lot of negotiation.

I've seen cases of top developers earning almost 50% less than their peers because they were hired via subcontractors.

> Average skill of LATAM devs is still low compared to the US

I'd say it is a sample bias of the developers working as contractors (average devs with good enough English).

Top LATAM talent moves abroad or works as employees at international companies that pay way more than nearshores.

I would love to read more about your experience. We need more content with feature, peformance, and architecture comparisons. Currently, there's a lot of developer evangelism hype in the space.

Yup. pgvector will do it for a lot of projects, specially if you're just trying things out. It think of it as using PostgreSQL full text search before you need to deploy a decidated solution.

I'm having success with simple YAML schema. One thing that's very helpful for the prompt is to include "description" and "example":

      -
         column name: salary_max
         format: number
         example: 150,000
         description: Salary Maximum
      -
         column name: keywords
         format: string
         example: engineer, python, docker, remote
         description: Relevant Keywords (Comma separated keywords used for filtering and matching jobs to candidates)

There's nothing to see here, just a waiting list.

Check Show HN rules:

> Off topic: blog posts, sign-up pages, newsletters, lists, and other reading material. Those can't be tried out, so can't be Show HNs. Make a regular submission instead.

https://news.ycombinator.com/showhn.html

This is looking nice. Would definitely try it if I was still doing SQL daily.

Small suggestion: show that it's a one time payment closer to the price. I kept staring at the price and thinking if it was yearly, until I looked at the button.

Thanks for releasing it. Loved the demo video.

I haven't heard of the Cloud API. Is it easier to get started with WhatsApp API now? Does it make services like 360dialog less relevant?

You've invested a lot of time on this! I doing something similar (completely automated) for another niche, let me know if you wanna exchange ideas (twitter in bio).