Nimsical

Unclear to me how much of the negative perspective of this is because of her gender. Media has a bias against female leaders. The fact that they don't put the whole email in there is concerning as well.

E.g. Stripe had a similar tone in that they basically wanted to maintain their margins but didn't get anything close to this type of negative coverage. No MLK quotes though

This is cool and in some world where the company becomes massive and basically commit to properly maintaining the product I could see it being a player.

I'm skeptical of open-sourcing UI and workflow builders.

The upsides are that you enable a community to build connectors and the UI builder + maintain them, but the downside is that you have to manage the community well enough that enterprises can trust the connectors and the UI builder. The challenge of maintaining the community + maintaining some sort of an SLA is very hard. This type of software is extremely hard to test for –writing integration tests are much harder for the frontend than they are for something like a database (e.g. MongoDB) because of the permutations of use-cases). The OSS+managed model seems to have succeeded in areas where very you can regression tests are much easier to maintain and there are clear benchmarks to test the community's output against.

As a buyer (we recently bought SuperBlocks (https://www.superblocks.com/) which is just a managed version of the same idea) it's hard to commit to an open-source version of something like this given the problems above. I may be totally wrong – I've never run an OSS+managed business, and as much as I love the ethos, I still have to not care about the stability of an internal tool builder that my company is being built on top of.

[Disclaimer: my company is using Superblocks and have been paying for it for about a year]

Been reading the comments and I think there are a bunch of questions about why companies would use a tool like this and that integrating your product / company into it is potentially risky. Here's why we decided to use it.

First thing you should know is that before we moved to Superblocks, we were using a combination of Retool, Google Cloud Scheduler, Mode Analytics, and a few small Google App Engine apps to solve our problems.

There is a whole ecosystem of apps and mini-products our engineering team has to build which require slow-paced maintenance, but are still mission critical to our business. Analytics dashboards, Slack reporters, revenue dashboards, customer service dashboards, Discord bots, etc.

There are plenty of different tools which solve some of those problems in silo, but the main problem (and beauty) of building internal apps is that you can fully solve for your unique business problem. We're a tech-enabled advertising agency (there are a handful of companies like ours in the world) and we both have our own external product as well as a series of internal apps that power our operations.

We decided to move all our internal apps to Superblocks effectively because we could merge a few different platform layer things (Google Cloud Scheduler, App Engine, Mode, Retool) and be able to build much faster within our team. The big thing for us was that we have engineers writing in Python (more data heavy stuff), and engineers writing Javascript. Previously, we had to have them offer APIs to each other and it wasn't as agile as it is today for simple things we had to build for our business stakeholders.

I remember seeing one of our first Workflows on Superblocks when our data scientist built a BigQuery step, a Python step, and then our Javascript engineers grabbing that to do some post-processing with it to display the data and thinking "this is probably how most tiny apps will get built in the future".

I don't want to have infrastructure for every tiny solution I built for our team. I just want things to work.

I've been using Superblocks (https://superblockshq.com/) to do this.

I'm generally a fan of reducing complexity for building internal tools. We used to use Retool with a combination of Google's App Engine, Google Cloud Scheduler, and Mode Analytics to support our internal teams but we consolidated all of it under Superblocks.

They have charts, you can build backend functions (and call them on a schedule!), and you can build a UI just like you can with Retool (they have feature parity in most areas with Retool on the UI side).

The best part has been that they support Javascript and Python at the same time, which has let our data and engineering teams live in the same eco system and work together to build internal data tools for our team members.

Huge fan!

Based on this data (someone else shared in another thread): https://data.firefox.com/dashboard/hardware – the numbers shared in the link on this post are misleading.

Looks like the sample size has some bias built into it given the data is from those who use their performance software. I would trust the numbers from Firefox much more (higher scale, much more ubiquitous product).

On 1: this is because they've done a great job with partnerships. Cerkly's corporation set-up becomes free if you sign up with FRB – this was a YC deal we used. So this doesn't mean much.

2. I recognize that this is generally a hard job. Our original product/company was along the same lines as Wirecutter (https://www.digitaltrends.com/mobile/ask-suto-launch/) and it's an extremely tough path to take. I do feel that it's easier for businesses since the value-props/features are a lot more distinct in how businesses make decisions.

Looking at your "The Best Store of Money" – I'm baffled that First Republic Bank even made the list. We're in the process of switching away from them because of how broken their web interfaces are, and how hard it is to connect them to other tools which makes it hard for our accounting team to automated their work.

I think you really need to do a lot more interviews with people that have been users of these products.

I've been a repeat startup employee, and have now become a founder. The reasons I chose to live this path haven't changed.

If you have a comp.-driven mindset (i.e. you'd like to become a millionaire in X years), then I think there are far faster, better, easier ways to achieve that. Join a top-tier firm in the valley / NYC and do a decent job at getting promotions and you'll be well within your way by your mid-late thirties.

The best part of being part of a startup is your ability to be in fast-paced multifaceted problem-solving environment.

You can play quiet a few roles at a small company. As an early employee (engineer or otherwise), you can look at any part of the company, product or its internal operations and decide that it's something you care enough about to fix and improve upon.

I joined my first startup as an engineer, became a product manager, did sales, built employee on-boarding flows and training manuals, re-branded the company and learnt how to manage a sizable team as I moved on to marketing. All within less than 2 years.

You can get most of those experiences elsewhere, but not within that timeframe or the depth of autonomy you would receive to get it done is not attainable in a Big Corp.

This is exceptionally great if you're early in your career, are generally good at getting tasks done, but have no idea what your calling is. You can work at a startup (or a series of them) for a few years and walk away with understanding what you'd like to focus your time and money on for the next few decades of your life.

Only caveat that I would add is that if you'd like to achieve any of that, then it's paramount that you look for the right founders. Founders that are willing to give away autonomy and let you flourish. They tend to be the best ones and produce great companies with incredibly rich cultures as a result.

This is just bad data journalism.

I clicked cause I thought it was about Star Trek. Always forget there's a library with that name.

I have no degree, but have had a documentary made about my first startup in Canada + speak at conferences / meet-ups from time to time, have a couple of pieces of press and have a profitable business with a handful of employees

Hey Peter, thanks for doing this :)

I'm on an O1, founder or a company and have slightly less than 50% onwernship.

I've been thinking of getting a green card. What's the best approach?

I'm on an O-1 Visa at the moment and thinking about going for a green card. What's the best path way to do this?

This is cool!

Wondering what you're using for OCR?

Siri's behind on a lot of stuff. They have issues gathering all parts of their organization and centralizing efforts for A.I. due to their org. structure, so it's going to be a while before Siri can catch.

On the pricing front – I'm not very surprised that there's a bump. They've always been on the high-end of everything they release (Apple TV is 2.5x-3.X other competitors from Google/Amazon).

I think the main reason they played down the "smart" part is because they know Siri is behind.

Sonos will have Alexa in it and most people can't tell the difference between a better speaker over another in terms of sound.

It just ...looks better.

There's no open API for this, right? And Siri is ~1-2 years behind Echo/Google Home. Only "innovation" is on aesthetic design.

Apple usually comes in when they know they can do something a lot better than the existing competition, but not recently. Not with Apple Watch & HomePod.

It's now about:

1) Apple has massive distribution 2) They can get a non-step-function revenue increment with that massive distribution, launching new products.

Which makes sense why it's music focused (lock-in with Apple Music)

There's no open API for this, right?

And Siri is ~1-2 years behind Echo/Google Home. Only "innovation" is on aesthetic design.

Apple usually comes in when they know they can do something a lot better than the existing competition, but not recently. Not with Apple Watch & HomePod.

It's now about:

1) Apple has massive distribution 2) They can get a non-step-function revenue increment with that massive distribution, launching new products.

Which makes sense why it's music focused (lock-in with Apple Music)

How is the clustering of these articles actually work?

Pretty interesting!

I don't track that information on the function – but I really should!

Based on StdLib's dashboards – a bunch of folks have been using it per month with a steady pace of a few 100 scrapes a day type of thing.

We've been using it internally for quite a while now.

And as far as I know, StdLib doesn't allow you to use your own AWS credentials. They have their own gateway and a bunch of stuff on top of Lambda that makes the whole experience a lot easier and more powerful (e.g. 128MB limit on payload vs 5MB for Lambda)

I tend to tell this to people quite often, but depending on the use-case – it might actually make sense to use something like Mechanical Turk to get complicated scraping done.

But if the content is pretty neatly and easily scraped then the other recommendations make sense.

Though a bit of Javascript can now take you a long way ;)

A ton of modern languages still use the namespace I think.

hah, the C/C++ programmer in me cringed the first time I saw it. It's a decent platform though.

I guess it's not truly distributed in that sense. StdLib uses AWS Lambda, which have widely different IPs and I believe they're multi-region.

I haven't had issues hitting a wall with getting caught doing any scraping. But then again, I haven't done it at a 10k/pages/sec rate or anything like that.

Mainly built this as an experiment to pull a bunch of data for some ML work I've been doing. Wrote about it more extensively here: https://hackernoon.com/microservice-series-scraper-ee970df3e...

AMA!

This is my favorite product in the world right now.

It's rather ironic that the front-end developer handbook doesn't support mobile devices.

Doesn't inspire confidence.

Was interesting to see a company take over a community open source initiative.

They don't always recommend products they make money off of. Plenty of their products (especially in the Sweethome version) are recommendations they won't make money off of.

They're probably the least biased reviewers out there at the moment and take it very seriously.

I've used Lambda in production under pretty high throughput conditions before. It's fairly stable – given that we were using Kinesis to feed into it (which was a pain to setup).

Recently, I've been using stdlib (https://stdlib.com) to do function-services since it's a lot easier to get started with it a bit more intuitive for newcomers to understand.

Kinda hard to trust company accounts with an external service like this. Is there a way to just give you a downloaded CSV or something for this?

If not – how are you managing security?