nreece

There was a related article on 'Turning satellite imagery into wall art', that had pictures of easing cloud cover from photos: https://news.ycombinator.com/item?id=28325421

>> They call the DB "relational" but I can't seem to find anything in the docs about relations, joins, cascade delete, etc.

Looks like they support table relationships via PK-FK (mentioned under Data Type section: https://catalyst.zoho.com/help/data-store.html#Column_Charac...), but there's no mention of querying or inserting relational data.

>> If you're selling to a team, don't. Sell to the person buying products/tools for the team, the one with signing authority.

While buying decisions are made on top, initial usage/trial decisions are often made by expert team members. I'll say, first soft-sell and get expert team members to try out your product, and soon-after ask them to connect you to the manager or decision-maker for further discussion.

Thanks for using and mentioning Feedity!

We've since rebranded to New Sloth: https://newsloth.com, which is now a simple integrated feed builder, reader and clusterer/deduplicator, specially aimed for knowledge workers with hundreds and thousands of feeds to monitor daily.

Besides visual selector-based feed generation, our API can auto-magically detect relevant selectors (in most cases) as well.

Here's a short video depicting some of these mudras as used in Bharatanatyam (a form of classical dance from Southern India): https://www.youtube.com/watch?v=I4FtwU2MVoE

Generally, POST is to create a new entity, while PUT is to fully update an existing entity. PATCH is used for partial updates, if necessary in a use case.

I was recently working with the Vultr API, and they follow this pattern as well: https://www.vultr.com/api/#section/Introduction

Dupe of https://news.ycombinator.com/item?id=27555013

Pretty cool! I'll try it out in more detail.

I also liked a similar product, aimed more at ETL/ML/AI segment, that was recently posted on HN: https://hal9.ai

Question: I added a 'HTTP Request' which returns JSON data, but connecting it to the 'Sort' block shows a warning/error: 'Data input is not valid. Accepted types are: DATASET'. So how do you convert JSON to a dataset?

Some more feedback:

* Explode module does not work in some cases. It shows no fields above the data preview.

* In between 'Import' and 'Visualize' steps on your website features section, you can think of having a 'Transform' step, which will sit well with ETL and ML folks.

* It'll be nice to auth via GitHub or just simple email (magic links), instead of Google only. GitHub auth will also allow you to support direct repo load and save for user's pipelines with change history.

'Pipeline running' delay is a pain, something potential users/customers won't wait around for. Have you looked at using on-demand managed containers (like Google Cloud Run) instead of full EC2s to launch the pipeline? It'll be cheaper, quicker and more scalable too.

Just saw that JSON support already exists, which is great.

Also, it has the potential of being a great platform, but I wouldn't suggest marketing it as a 'Roblox for AI'. How many AI/ML people really use Roblox? Few I suppose. Perhaps something simpler and familiar, like 'GUI for AI' or 'Your ML Lab/Studio/Toolbox'.

Played with it a bit. Looks good, although rough around the edges as expected of an alpha.

Lot of times, it gets stuck at 'Pipeline running', specially with dataset and web tables. Ability to fetch remote data via web JSON and XML/RSS will be nice.

Correct link (PDF): https://fosshost.org/Fosshost_Donations.pdf

Google Cloud Run will the quickest to start with, most scalable and least expensive in the short run.

Vultr is all you'll need for inexpensive and reliable VMs. My experience with Vultr has been much better than Linode and DigitalOcean. RamNode is also great in terms of performance.

Been hearing good things about Render's unified cloud too. Worth checking out.

Caveat: It is estimated that 70% of wealthy families will lose their wealth by the second generation, and 90% will lose it by the third.

https://www.nasdaq.com/articles/generational-wealth%3A-why-d...

https://money.com/rich-families-lose-wealth/

Speaking of runtime support, any reason why C# is using Mono instead of official cross-platform .NET Core 3.x or the latest .NET 5, and it also looks like ASP.NET is unsupported. Any plans to address these in the near future?

Marketing and positioning definitely needs a push towards robust and scalable hosting.

Integrated environments (browser IDE + CI/CD + serverless hosting) are going to be big for starter apps and quick APIs.

I spent some time reviewing Google Cloud Shell Editor, which has git support (not GitHub yet I guess) and one-click build & deployment to Google Cloud Run, which overall is a powerful integrated environment in itself.

Imo, replit should preemptively focus on improving and highlighting hosting.

Curious to know if anyone's hosting real-world production web apps or APIs on replit, or is it mainly an educational platform?

In my opinion, and experience from my own startup journey, a 'product-market fit' is simply being profitable from revenue that's growing quarter on quarter for at-least an year or two.

As as aside, I recently goofed up our company website DNS (updated a record on long TTL with an incorrect), but quickly fixed it and found a partial workaround to propagation: flush the DNS cache of Google[1] and Cloudflare[2]. It helped with DNS cache refresh within minutes from most global locations, if not all.

[1] https://dns.google/cache [2] https://1.1.1.1/purge-cache/

RSS is dope. At Feedity, we love how feeds make it immensely easy and valuable to monitor hundreds and thousands of news sites, insights, blogs, jobs, products and what not.

With our rebranded and revamped platform, New Sloth (https://newsloth.com), we continue to generate feeds for public sources and aggregate content across channels, all neatly collated in one place, making it super useful for busy professionals and savvy businesses to stay ahead of the crowd and make informed decisions.

I wanted to support and use Plausible, mainly for the drill-down feature, which Fathom doesn't have yet.

But, I had two main concerns about this product from my trial:

1) Slow response time (~300ms) of the script when using a custom domain.

2) I liked its simple design and features, but looks like it's already starting to bloat up with trivial things like UTM tracking.

Some small UI bugs (like time range buttons not working) also need ironing out.

Waiting for Cloudflare Web Analytics to compare.

Good stuff Pieter. Simple is good.

It'll be interesting to see your code organisation. Something to learn from, as my single-page projects often get cluttered and hard to manage or add features.

Btw, for some strange reason you have me (@voidmonk) blocked on Twitter :) so I can't see your tweets.

An existing community (like Medium or Substack) can help gain readership, else you're on your own to promote your content.

I recently came across a popular new blogging platform - Hashnode. It has a decent balance between ownership (sync to repo) and distribution (their community).

The real benefit seems to be its in-built CI/CD, and horizontal scaling, much like GCP Cloud Run:

"App Platform allows you to hook up your source code repo and deploy using a cloud-hosted build system that automatically detects the needed environment to run your app."

We had a similar experience, so I'm just sharing the path we took.

We were using New Relic APM and Server Monitoring through the Microsoft BizSpark benefits for an year or two. We were even ready to pay for their licenses, but New Relic pricing for multi-server use was starkly high for a small tech company like ours. Their product was good, but their focus was more on enterprise customers.

After trying out many other commercial products (focusing on .NET Core stack in particular), and keeping long-term scalability and TCO in mind, we decided to utilize some open source solutions.

First we tried Bosun[1] (from the Stack Exchange team), but found little progress being made on it. Then, we tried Prometheus[2] collectors for monitoring, Grafana[3] for the dashboards and alerting, and Exceptional[4] for logging. One person got it all working within a week. We've never looked back, and have to hardly touch the set-up anymore (except for system updates once in a while). It has worked beautifully for almost two years now, going from 2 to a few dozen apps/servers/VMs.

Hope it helps you, and others, explore the open source monitoring route.

[1] https://bosun.org [2] https://prometheus.io [3] https://grafana.com [4] https://github.com/NickCraver/StackExchange.Exceptional

Vultr Object Storage (S3 compatible): https://www.vultr.com/products/object-storage/

Backblaze B2 Cloud Storage (S3 compatible): https://www.backblaze.com/b2/cloud-storage.html

This looks great!

I read your blog post about how open-sourcing helped you find testers, which is a great step for such tools.

* Shameless plug *: Our web service, Feedity - https://feedity.com, helps create a custom feed for any news webpage, via a point-and-select feed builder and REST API.

Algo is easy and works well: https://github.com/trailofbits/algo

Been hearing the same hype about Roam Research, just as Slack, Zoom, and countless other 'shiny new' consumer tech products.

Your answer lies in your question. What would you pay a small amount of money for? Go build that.