Hacker News: tuckerconnelly

New comment by tuckerconnelly in "YuE2 · Frontier Music with Symbolic Planning"

tuckerconnelly — Fri, 11 Sep 2026 01:55:05 +0000

Much like the restaurant industry, creative industries will probably bifurcate. Cheap (AI, drivethrough, takeout) for daily and most use cases, experiential (vinyl records, concerts, upscale restaurants) for special occasions, connoisseurs.

New comment by tuckerconnelly in "Ask HN: How are you finding the job market in July 2024?"

tuckerconnelly — Thu, 01 Aug 2024 02:03:59 +0000

Hey I run a little recruiting startup, and we have a few roles at fintech companies in NYC--wanna send me your LinkedIn + resume?

New comment by tuckerconnelly in "Why is Chile so long?"

tuckerconnelly — Tue, 02 Jul 2024 19:23:25 +0000

Yeah based on my in-laws, Chilean seems understandable, though they tend to speak fast.

Argentine Spanish is the strange one, due to the heavy Italian influence.

New comment by tuckerconnelly in "Ask HN: Why do message queue-based architectures seem less popular now?"

tuckerconnelly — Wed, 19 Jun 2024 00:44:33 +0000

I can offer one data point. This is from purely startup-based experience (seed to Series A).

A while ago I moved from microservices to monolith because they were too complicated and had a lot of duplicated code. Without microservices there's less need for a message queue.

For async stuff, I used RabbitMQ for one project, but it just felt...old and over-architected? And a lot of the tooling around it (celery) just wasn't as good as the modern stuff built around redis (bullmq).

For multi-step, DAG-style processes, I prefer to KISS and just do that all in a single, large job if I can, or break it into a small number of jobs.

If I REALLY needed a DAG thing, there are tools out there that are specifically built for that (Airflow). But I hear they're difficult to debug issues in, so would avoid at most costs.

I have run into scaling issues with redis, because their multi-node architectures are just ridiculously over-complicated, and so I stick with single-node. But sharding by hand is fine for me, and works well.

New comment by tuckerconnelly in "OpenAI Selects Oracle Cloud Infrastructure to Extend Microsoft Azure AI Platform"

tuckerconnelly — Tue, 11 Jun 2024 22:07:12 +0000

What are your gripes with GCP? I've been using it for every project for a while, and am super happy with it, especially GKE.

New comment by tuckerconnelly in "Jobs Aren't Coming Back"

tuckerconnelly — Tue, 04 Jun 2024 17:14:45 +0000

You don't see yourself hiring help for:

* Keeping things up and running

* Building the systems that the AI interfaces with

* Supporting any large contracts

In my experience, one enterprise contract with specific requirements basically requires you to hire a dev to support them.

New comment by tuckerconnelly in "Jobs Aren't Coming Back"

tuckerconnelly — Tue, 04 Jun 2024 17:09:05 +0000

My counterpoint: if your product is simply a ChatGPT wrapper, you have no moat. Whatever is complicated enough and is actually making money, and you feel the need to test it and make sure it keeps running, that's your moat, and that's what you're going to want to hire human help for once you actually make some money.

New comment by tuckerconnelly in "50% discount on OpenAI pricing if you submit a batch and give them up to 24h"

tuckerconnelly — Mon, 15 Apr 2024 18:51:30 +0000

Message queues are difficult to maintain.

Webhooks is the best solution IMO.

New comment by tuckerconnelly in "Ask HN: Is anybody getting value from AI Agents? How so?"

tuckerconnelly — Sun, 31 Mar 2024 23:43:59 +0000

There's value, but it's too expensive, too slow, and too unreliable right now to be feasible from a business perspective.

New comment by tuckerconnelly in "Ask HN: Is anybody getting value from AI Agents? How so?"

tuckerconnelly — Sun, 31 Mar 2024 19:33:00 +0000

Chain of thought takes time to generate all the characters. If you do a chain-of-thought for every action and every misstep (and you need to for quality + reliability), it adds up.

New comment by tuckerconnelly in "Ask HN: Is anybody getting value from AI Agents? How so?"

tuckerconnelly — Sun, 31 Mar 2024 18:03:20 +0000

I built an AI-agents tech demo[1], and am now pivoting. A few thoughts:

* I was able to make a simple AI agent that could control my Spotify account, and make playlists based on its world knowledge (rather than Spotify recommendation algos), which was really cool. I used it pretty frequently to guide Spotify into my music tastes, and would say I got value out of it.

* GPT-4 worked quite well actually, GPT-3.5 worked maybe 80% of the time. Mixtral did not work at all, aside from needing hacks/workarounds to get function-calling working in the first place.

* It was very slow and VERY expensive. Needing CoT was a limitation. Could easily rack up $30/day just testing it.

My overall takeaway: it's too early: too expensive, too slow, too unreliable. Unless you somehow have a breakthrough with a custom model.

From the marketing side, people just don't "get it." I've since niched down, and it's very, very promising from a business perspective.

[1] https://konos.ai

New comment by tuckerconnelly in "Our Company Is Doing So Well That You're All Fired"

tuckerconnelly — Thu, 22 Feb 2024 20:59:56 +0000

While X's main use case works, I just tried to run some ads there.

When trying to upgrade to verified org: "An unexpected error occurred"

When trying to change my display name: "An unexpected error occurred"

When trying to run an ad: "Awaiting verification" for a week

Contacting support through messages: No response

Finally finding a way to submit a ticket: 3 day response time with a standard response that didn't resolve the issue.

Responded to that ticket and haven't gotten a response in days.

I don't think they're operating at the same level as they used to :)

Pet Safe Kennels

tuckerconnelly — Mon, 29 Jan 2024 04:02:05 +0000

Article URL: https://www.pet-safe-kennels.com/

Comments URL: https://news.ycombinator.com/item?id=39172552

Points: 1

# Comments: 1

New comment by tuckerconnelly in "Discord is laying off 17 percent of employees"

tuckerconnelly — Thu, 11 Jan 2024 19:50:48 +0000

How does 0% interest rate actually affect this though? Are that many companies actually funded on debt now? Or are are sales down because their customers were purchasing with debt?

New comment by tuckerconnelly in "Disney's earliest Mickey Mouse enters public domain as US copyright expires"

tuckerconnelly — Mon, 01 Jan 2024 17:17:19 +0000

Cuphead should absolutely add a Mickey Mouse DLC.

New comment by tuckerconnelly in "Cleveland launches plan to provide cheap broadband"

tuckerconnelly — Tue, 24 Oct 2023 01:46:36 +0000

I just visited a few days for a surgery and this seems pretty accurate.

The food weirdly was the same at every restaurant, regardless of the cuisine: super oiled up. Usually mayonnaise and something pickled on top. Like preparing me to go to a factory for 14 hours.

Nearly every Uber had a sports station on.

Everyone seemed pretty gloomy there. "What's a must-do before we leave?" "Umm...yeah it's too cold for a Brown's game, so not much. Yeah."

A strange obsession with Taylor Swift "bringing money into the city" because of the sports guy she's dating now from Cleveland.

Ask HN: UI for Managing Celery Tasks?

tuckerconnelly — Fri, 20 Oct 2023 19:10:16 +0000

I'm used to bull-queue, which has excellent UIs for managing jobs (retry, delete, etc.) but python support is limited at the moment.

Flower is really bad IMO for managing jobs.

Is there an existing solution for managing celery tasks (paid or open source)?

Or if I really had to build one, if I built it, would your company pay for it?

Comments URL: https://news.ycombinator.com/item?id=37960195

Points: 1

# Comments: 0

New comment by tuckerconnelly in "Roundup herbicide ingredient connected to epidemic chronic kidney disease"

tuckerconnelly — Mon, 16 Oct 2023 20:39:43 +0000

> But when glyphosate encounters certain trace metal ions that make water hard—like magnesium and calcium—glyphosate-metal ion complexes can form. Those complexes can persist up to seven years in water and 22 years in soil.

Isn't basically all water outside hard water? How did public health agencies miss this?

New comment by tuckerconnelly in "Ask HN: Bad Results from Vector Search?"

tuckerconnelly — Fri, 06 Oct 2023 02:33:19 +0000

Thank you! I've used the CrossEncoder from the sbert library you mentioned to rank the vector search and keyword search results, but it still doesn't help that the vector search results are odd.

This video suggests Elasticsearch is (or will be) able to blend the two: https://www.youtube.com/watch?v=5Qaxz2e2dVg

I'm hoping meilisearch adds that because ES is a beast of a software package.

Ask HN: Bad Results from Vector Search?

tuckerconnelly — Fri, 06 Oct 2023 01:30:44 +0000

Hi, I'm building an RAG system with OpenAI embeddings and the ChatGPT api.

I chunked all the documents into 400-800 character chunks, vectorized them all, and put them in a vector database.

The results are pretty bad--the surfaced document chunks kind-of-but-not-really match up with the query.

I'm getting much better results from simple keyword searches (using meilisearch).

Am I doing something wrong? Do I need to use a fine-tuned model like BERT? Is this technology vastly overhyped?

Comments URL: https://news.ycombinator.com/item?id=37786273

Points: 2

# Comments: 5