How well agents can do this is mostly proportional to how well they can understand and navigate your codebase broadly.

There are various contributing factors to this, but they include clear docs, notes and refactors that clear up parts the agent commonly gets confused by, choosing boring technology (your dependencies are well understood) and access to command-line tools that let it lint + typecheck + test the code. A lot of the scaffolding and wiring necessary are built into Cursor and Claude Code themselves now. Hope that helps!

This is a great technique!

In this case, I had made an overlarge squashed merge that included both the Intercom integration (a suspiciously likely cause of slowness) and the feedback button that added the heart – so I needed to go deeper to figure out the true cause. (Noto Emoji was in the app from before, but wasn't triggered in the dashboard until we added an emoji there.)

Not off topic at all!

While in this case we’d included the emoji font for displaying user content in another part of the app, the hazard of letting a “simple” approach expand and get out of hand is part of what I wanted to convey in writing this.

The FDA (or equivalent in the relevant country) regulates whether an approved drug requires a prescription based on the safety profile. To be approved for OTC, there is a much higher bar in terms of ease of misuse, side effects, and so on.

My experience raising in Vancouver is that there are two angel communities.

There is an easier to find Vancouver-centric investor group that behaves as a you describe. Many of these investors didn’t come up as tech founders. I was advised not to waste time with them, so I don't know if there are some gems in the rough there or not.

Then there is a quieter group that got their capital from building serious tech businesses. This group spends more time connecting outside Vancouver – Bay Area mostly, but also Toronto and globally. These folks do write early-stage cheques and can be very helpful advisors, but they're not full-time angels who are spending time on deal flow. They're mostly focused on building their next thing, so it’s more difficult to earn their attention.

So yes there are questionable actors, but there are also very legit folks doing great work, and it’s possible to go to an event or dinner party that only really has one or the other. Hope that’s helpful to any other founders building in Vancouver!

Yes, well said. This is what I noticed as well, though it me a few years to work it out.

Thanks Grey – other than the presenting-at-an-event flow I do really did like the Figma Slides experience, so this is great to hear. The world is better off with a strong Figma.

This is interesting, but the values could use some refinement:

  gpt-4o: speed 3/5
  gpt-4.5-preview: speed 3/5
  gpt-3.5-turbo: speed 2/5

In practice 3.5 turbo is, what, 5-10x faster than 4.5 preview?

I mentioned Ruby and Python in this vein, but PHP with Laravel would count in my book too. That said, PHP may be nearing the stage Java has gotten to, where it’s well-understood but is perceived as outdated enough that it causes hiring and retention friction.

While this can be done in principle (it's not a foolproof enough method to, for example, ensure an LLM doesn't leak secrets) it is much harder to fool the supervisor than the generator because:

1. You can't get output from the supervisor, other than the binary enforcement action of shutting you down (it can't leak its instructions)

2. The supervisor can judge the conversation on the merits of the most recent turns, since it doesn't need to produce a response that respects the full history (you can't lead the supervisor step by step into the wilderness)

3. LLMs, like humans, are generally better at judging good output than generating good output

Yeah this is an interesting point. Other threads make the point about the "bitter lesson", and how expert-trained ML has historically not scaled, and human-generated LLM training data may just be repeating that dead end. Maybe so.

Something that is new this time around, AFAIK, is that we haven’t previously had general ML systems that businesses and consumers are paying billions of dollars a year to use. So if, say, 10% of revenue goes back in to making better data sets every year, I can imagine continued improvement on certain economically valuable use cases – though likely with diminishing returns.