Hacker News Reader

I agree that some GCP services are better than others.

I’ve never used Pub/Sub or Cloud Run, but have been quite happy with BigQuery and GKE.

rdlion Ask HN: What are good books/blogs to read for a first time CTO?

I think the #1 thing when you become part of an exec team is that you should be optimizing for the _business_, and not your function. The working assumption is that you will keep your function executing and delivering, but what is really hard is helping to figure out what the right decisions are for the business. Should we invest more in product or sales? What if there’s a huge top of funnel problem — what can we do about it? Your job is to bring that technology perspective to the discussion.

I’ve been an exec, founder, CEO, and board member at various stages of successful (IPO) /unsuccessful companies (acqui-hire) companies. And the common thread at every stage is that the most successful companies had management teams that worked well together to optimize for the business.

So instead of spending your energy on reading / learning more about tech, I’d recommend you spend your energy learning more about business (I’d probably start by asking the CEO & the rest of the mgmt team for advice on what to learn.)

rdlion Ask HN: SaaS Founders, What 3 advice would you give your younger selves?

At $1M ARR, I think a CFO is overkill (and hideously expensive). In my experience:

- $0 - $1M-ish, hire a part-time bookkeeper

- $1M - $5M-ish, get a director of finance / controller, who can manage AR/AP, audit, ASC 606, basic financial modeling, and your key rev benchmarks.

- $5M+ is when I'd consider hiring a CFO, but it really depends on your growth rate. You can probably get to $10M+ with a good dir of finance if you get them to hire a good FP&A person.

rdlion Prophet: Automatic Forecasting Procedure

A common strategy is interpolation. The challenge is that forecasting itself is a form of interpolation. So you're forecasting based on forecasted data.

rdlion Prophet: Automatic Forecasting Procedure

I tried Prophet via Darts, and all the models in Darts assume a regular time series.

Re: "fancier machine learning" -- I've seen different flavors of RNNs & LSTMs have some success in analyzing time series data. I've struggled to get them to work on real-world (i.e., messy) data, but have had some encouraging results with a transformer encoder-only NN.

rdlion Prophet: Automatic Forecasting Procedure

As others have pointed out, Prophet is not a particularly good model for forecasting, and has been superseded by a multitude of other models. If you want to do time series forecasting, I'd recommend using Darts: https://github.com/unit8co/darts. Darts implements a wide range of models and is fairly easy to use.

The problem with time series forecasting in general is that they make a lot of assumptions on the shape of your data, and you'll find you're spending a lot of time figuring out mutating your data. For example, they expect that your data comes at a very regular interval. This is fine if it's, say, the data from a weather station. This doesn't work well in clinical settings (imagine a patient admitted into the ER -- there is a burst of data, followed by no data).

That said, there's some interesting stuff out there that I've been experimenting with that seems to be more tolerant of irregular time series and can be quite useful. If you're interested in exchanging ideas, drop me a line (email in my profile).

rdlion Early Days of AI

Another parallel that I see is “cloud”. Salesforce was one of the first (if not the first) real successful cloud companies. But it took a really long time (and it still is going on) for cloud-first to be the default in enterprises. And along the way, there was a ton of new technologies that were (re)invented: VMs and then containers etc at the infra level, new app architectures like AJAX (holy smokes was this amazing when it first came out), etc.

Similarly I think we’re in for a wild ride on AI and figuring out its implications. There’s a ton of obvious use cases today but I’m really interested in the ones that aren’t obvious right now.

rdlion Ask HN: If we train an LLM with “data” instead of “language” tokens

Temporal Fusion Transformer: see https://arxiv.org/abs/1912.09363.

rdlion Langchain Is Pointless

I found it helpful for prototyping and learning some basics, but I quickly found the abstractions were not useful and had to implement my own.

rdlion Ask HN: Could you share your personal blog here?

https://www.thelis.org/blog

Mostly notes to myself, and I’ve got a half-baked post on LLMs that I’ve been working on (as that’s been my area of recent interest).

rdlion Ask HN: Who wants to be hired? (July 2023)

Location: Boston, MA

Remote: Yes, or in Boston

Willing to relocate: No

Technologies: Kubernetes, Envoy, LLMs, service mesh, Hugging Face, Python, SQL, BigQuery, DBT, LangChain, Docker

Résumé/CV: https://www.linkedin.com/in/richardli/

Email: rdl@amorphousdata.com

I posted last month and had a few worthwhile conversations, so posting this again!

I'm an entrepreneur & product person. I've been an exec at two successful exits (Duo, Rapid7) and started my own company (Ambassador Labs) whose story is still being written. I'm super-interested in LLMs and what issues people are running into. So I'm interested in anyone who could use a product person for a part-time consulting gig. The more interesting the consulting, the less you need to pay me! :)

Some examples of work I could do would be helping you tell your story, building a sales deck, or maybe figuring out what you could do with all the new AI stuff. Or something else. Thanks for reading.

rdlion 3M heads to trial in ‘existential’ $143B forever-chemicals litigation

There's definite leaching of plastic compounds into food, which gets exacerbated when heated. My concern is the number of unknown unknowns. BPA became a big part of the consciousness a few years ago, and now it's PFAs, but what else?

https://www.theguardian.com/us-news/2020/feb/18/are-plastic-...

My general view is that glass is super-durable, microwave-safe (I would never microwave Tupperware), and the cost tradeoff is minor, so it seems worthwhile. That said, if I order takeout and it comes in a plastic container that's hot ... I still eat it :).

rdlion 3M heads to trial in ‘existential’ $143B forever-chemicals litigation

I don't think Scanpan is PFA free: https://www.scanpan.com/chemical-components.

rdlion 3M heads to trial in ‘existential’ $143B forever-chemicals litigation

We've worked to reduce plastics & chemicals in our house. Some things we do:

* We use silk dental floss (we use Radius)

* We use glass storage containers instead of Tupperware

* For cooking, we use All-Clad.

* If a recipe calls for non-stick (e.g., pancakes) I use a braiser from Le Creuset, which works reasonably well.

(Edited: formatting)

rdlion Ask HN: Who wants to be hired? (June 2023)

Location: Boston, MA Remote: Yes, or in Boston Willing to relocate: No Technologies: Kubernetes, Envoy, LLMs, service mesh, Hugging Face, Python, SQL, BigQuery, DBT, LangChain, Docker Résumé/CV: https://www.linkedin.com/in/richardli/ Email: rdl@amorphousdata.com

So this is a bit of an odd post, but I thought I'd try this out!

Some examples of work I could do would be helping you tell your story, building a sales deck, or maybe figuring out what you could do with all the new AI stuff. Or something else. Thanks for reading.

rdlion AI Canon

I was an early member of the CNCF community (circa 2016), and at the time I thought "wow things are moving quickly." Lots of different tech was being introduced to solve similar problems -- I distinctly remember multiple ways of templating K8S YAML :-).

Now that I'm spending time learning AI, it feels the same -- but the innovation pace feels at least 10x faster than the evolution of the cloud native ecosystem.

At this point, there's a reasonable degree of convergence around the core abstractions you should start with in the cloud-native world, and an article written today on this would probably be fine a year from now. I doubt this is the case in AI.

(Caveat: I've only been learning about the space for about 4 weeks, so maybe it's just me!)

rdlion Product Market Fit

In my experience, the founder(s) are one of the few people who will be able to push and innovate as others start to copy you. I think PMF is a continuum and what is PMF for $1M ARR is highly unlikely to be sufficient to get you to $10M ARR which will require more evolution to get to $25M ARR.

rdlion I surveyed 500 startup founders about their salaries

In the short-term this might work, but if you're trying to build a company that grows, then it doesn't work so well, because other people you hire will ... want the same thing.

If you expense your cell phone bill every month, then it's reasonable for other employees to expect the same, if they're also using it for work.

Better to get the board to sign off on a slightly higher salary to offset these expenses, IMO.

rdlion I surveyed 500 startup founders about their salaries

You'd need board approval.

rdlion Bridge Loans

I think the standard is fully convertible (i.e., you get your liquidation preference OR you convert) — a 1x as I think about it is you get your liquidation preference before you convert (i.e., two bites of the apple). Is that what you mean?

rdlion Bridge Loans

On a typical convertible note, I’d expect it to convert to equity on the same liquidation preference as the next round, and I do think even in this (bear) market I still think fully convertible deals are the norm.

rdlion Minikube quickly sets up a local Kubernetes cluster on macOS, Linux, and Windows

(Note: I work on Telepresence)

There's a bunch of semi-opinionated tools like Skaffold: Garden, Tilt (just bought by Docker), Docker Compose, and of course, the trusty ol' shell script powered by kustomize.

We wrote Telepresence because we found that everyone has their own snowflake-ish workflow once you get beyond the base case of "build container, kubectl apply", so we decided to focus just on the inner dev loop, and then integrate with your preferred workflow. YMMV.

rdlion The Future of Kubernetes

> The flip side of community ownership is more politics, more complexity, bureaucratic decisionmaking, and slower evolution.

This of course may be true for some communities, but certainly not the case for L7 proxies. In fact, at the time, the opposite was true: NGINX & HAProxy were content in slowly supporting their (captive) customer bases, with little incentive to innovate. This is why Lyft went off to build Envoy Proxy. At GA, it supported features such as HTTP/2, native observability, hitless reloads — none of which were easy / possible to do in NGINX or HAProxy at the time.

> Connection/service proxies are a dime a dozen. I suspect you could have picked nginx, haproxy, or the Linkerd proxy and been just as well off.

Linkerd proxy didn’t exist then (there was an early version of it, written in Scala IIRC; the Rust version didn’t come along to quite some time later).

I agree that there are many options for service proxies. I’ve found, though, that the Envoy community has been pushing innovation in the space quite aggressively.

rdlion The Future of Kubernetes

Envoy was originally written by Matt Klein @ Lyft, and the IP is owned by the Cloud Native Computing Foundation.

Here's the initial video where Matt Klein introduces Envoy: https://www.microservices.com/talks/lyfts-envoy-monolith-ser....

If you want to buy commercial support for Envoy & Istio, Tetrate certainly is an option.

(Note: I am/was the CEO at Ambassador Labs, which organized the Microservices Practitioner Summit where this talk occurred.)

rdlion The Future of Kubernetes

Istio was started by Google & IBM, while Envoy was created by Matt Klein at Lyft.

One of the reasons why we adopted Envoy at Ambassador for our API Gateway was precisely because there was no single _commercial_ entity who controlled Envoy.

As a note, technically Google & Lyft are VC-funded, so that’s true, although I wouldn’t put Google in the startup category, and Lyft’s business model is unrelated to Envoy.

rdlion Be curious, not judgmental

In improv, there’s an exercise where you say “yes and”. This language construct helps you build on what someone else is saying.

(Because in improv, if you object to something, you can just stop the show.)

It works really well in a professional setting, too. Try it!

rdlion Rivian IPO Roadshow Video

I love watching these videos; if there is an enterprising programmer who could set up a system to download these videos (or at least email alerts when new ones are published), I'd buy you a cool Lego set :-D!

rdlion The Sequoia Fund: Patient capital for building enduring companies

With the amount of capital available, no question that many companies that would have IPO'd years ago are electing to stay private (hello, Stripe!).

The cynic in me says: this is great marketing, because the best Sequoia fund is the early stage fund (always oversubscribed) and now as an LP you can't invest in just the early stage fund. (But the reality is I bet they forced early stage fund LPs to invest in their growth vehicles anyway, so there is no material change other than marketing. Which is very good.)

rdlion Facing sky-high connection fees, rural Ontarians go off the grid

I've been interested in how to do this -- in particular, how you tie in your solar inverter, windmill, and batteries together. Are you using AC to tie everything together, or are you able to use DC everywhere until you use it?

rdlion Understanding Startup Offers

The post says "an alternative career accelerated through learning, wealth, and reputation" ... and then doesn't talk about anything other than equity.

The article says "Equity will be your largest driver of compensation at a startup." as the rationalization of why it focuses on that.

Based on my past experiences, I would say that the learning, network, and reputational effects resulted in far more wealth to me over the medium-term than any incremental change in equity or salary.