1) Pre-PMF rule: Always be < 2 weeks away from a ground up rewrite.
Pre-PMF, most of what you build won't work & once you hit PMF you can't pause the work —> so, you carry over junk from earlier.
Don't accumulate debt in your project management!
1/n
2) Post-PMF rule: Don't depend on anything that is not in production!
Teams tend to make projects depend on each other to avoid duplication and maximize efficiency — but it's better to duplicate something than to depend on what is not yet in production.
2/n
3) Tech for non-tech founders: Make the effort to understand the tech canvas - the same way you understand the effect of inflation on your fundraise w/o an Economics degree!
You don't need to learn to code but you need to know the constraints of your tech and how it moves.
3/n
4) Hiring: Titles are designed for recruiters, not for executors.
An "Android engineer" isn't someone who knows Android *really well*, it's someone who *only* knows Android. You want the former, not the latter.
Optimize for execution, not for ease of hiring!
4/n
5) Hiring:
- Don't do artistic interviews (i.e. create on the spot questions). Instead, use standardized questions
- Use 2 interviewers —> prevent biases
- Successful hire = someone still on the job & doing well 6 months in
6) Network driven hiring: Build a network with people you want to hire
@ponnappa says his hiring pipeline is yrs long. @Twitter / @Clubhouse / @newsycombinator are great funnels. Build relationships early on so you know how they think. Good hires can be a year long process!
6/n
7) Generalists vs specialists: At <50 engineers, don't need specialists. By 500 engineers you have specialized teams. Somewhere in the middle you hire specialists.
Don't need QAs. Your best engineers should be flagging the issues themselves.
- Have a high bar & fire people who don't meet it. A bad engineer causes more stress for good engineers. Maintaining a bar creates a better work life for others.
- Don't prevent people from leaving. Do 1:1s so you always know in advance when they're leaving.
8/n
9) Common terms: Create a glossary of terms that lives across the entire company - everyone refers to the same glossary.
Every team uses different terms to refer to things➡️build a shared understanding of labels. Critical to dealing with invisible complex systems.
9/n
10) Cadences: Control all execution through cadences. It's the founder's job to create this heartbeat.
- End of each cadence: something needs to be in production
- Longer cadence➡️less clarity
- Have a documented list of all cadences running
- Cadence should be < 1 week
10/n
11) Involvement of leadership in the coding: Up to 100 tech people - everyone should be hands on.
- Manager / Tech Lead (Level 1) - Be fully hands on
- Manager of manager / VP (Level 2) - Be partially hands on
- SVP (Level 3) - Should ideally be involved with the code
11/n
12) Writing: Need to make your engineers comfortable with writing & communication - it's the barrier to good documentation
@ponnappa runs daily drills in bootcamps:
- Write a 2-3 para piece on what was done & publish it publicly
- Pick a random topic & just talk on it
12/n
13) Finally, @ponnappa's recommended reads on tech org building:
- Hackers and Painters by Paul Graham
- Mythical Man-Month by Fred Brooks
- No Silver Bullet by Fred Brooks
- The Gervais Principle by Venkatesh Rao
In Part 2 of @BlumeVentures' Edtech series, we deep dive into the ‘Continuous Learning’ sector. This includes any learning activity after formal education (school, college, & testprep) is complete.
As Twitter last week became the 1st co to announce permanent remote work, several 1st/2nd order effects were assessed (SF real estate $, 🚗traffic, rental $ in Tier2 cities, etc). Now as Square, Facebook, Shopify follow suit, it would be interesting to see other effects👇
1/10
1a/ Temp workers outnumber full time workers in several tech cos inc the likes of Google. Several temp workers are blue collar workers (cooks, janitors, guards) at massive silicon valley campuses. As WFH becomes permanent, demand for such blue collar workers will decrease
2/10
1b/ Not surprisingly, minority groups have a disproportionately large representation here- while 10% of tech workers in SF are Black/Latino, ~60% of blue collar workers in tech cos are Black/Latino➡️high risk of widening an already wide income gap
(siliconvalleyrising.org/files/TechsInv…)
3/10