5 ms·

"Move fast and break stuff" is a decent philosophy if and only if the consequences of breaking stuff are survivable. If breaking stuff means that your website

by dsr_ 8y ago

"Move fast and break stuff" is a decent philosophy if and only if the consequences of breaking stuff are survivable. If breaking stuff means that your website looks weird, that's survivable. If breaking stuff means that performance sucks for a while, that's survivable. If breaking stuff causes unavailability during a critical period of end-user demand, a few incidents might be survivable. If breaking stuff causes your company to have a terrible reputation for privacy, security, or competency, that might not be survivable. If breaking stuff causes your company to divulge financial information, that might not be survivable. If breaking stuff ends up costing your customers any significant amount of money, that probably will not be survivable.

credit_guy 8y ago
If breaking stuff causes a pedestrian to be killed because "they came from the shadows"[1], that appears to be perfectly survivable... /s [1] https://en.wikipedia.org/wiki/Death_of_Elaine_Herzberg https://en.wikipedia.org/wiki/Death_of_Elaine_Herzberg
- ttty 8y ago
  Cars killed people... they didn't stop creating cars. Nor would be a good idea.
  - abraham_lincoln 8y ago
    No, we got jaywalking laws and our rights reduced.
    - anticensor 8y ago
      We got safer cars at the same time. It is a double-edged sword.
  - TheCoelacanth 8y ago
    Killing one person per 1 million miles driven like Uber's self-driving cars have done is drastically different from killing one person per 80 million miles like human drivers do.
  - drngdds 8y ago
    It would be a great idea if there was a good alternative in low population density areas.
dralley 8y ago
e.g. Knight Capital. Let's not also forget the consequences of "breaking stuff" when applied to industrial process control, avionics, car "autopilot" (cough) etc. "Survivable" is sometimes a literal term.
ryanlol 8y ago
>If breaking stuff causes your company to have a terrible reputation for privacy, security, or competency, that might not be survivable. I reckon this stuff is particularly survivable. You can be like Linode, fuck up everything over and over again and still do just fine.
- osrec 8y ago
  Or Facebook...
- coldtea 8y ago
  >You can be like Linode, fuck up everything over and over again and still do just fine. Source? Because every time I've read about Linode they had glowing user reviews, including on HN...
  - detaro 8y ago
    a starting point (surprisingly high result on Google while looking for a list) https://news.ycombinator.com/item?id=10845278 https://news.ycombinator.com/item?id=10845278
  - freehunter 8y ago
    It blows my mind that people still give glowing reviews to Linode when in 2013 hackers breached their systems and stole credit card numbers and customers reported their credit cards were being used maliciously and Linode got away with saying "nuh uh": https://www.theregister.co.uk/2013/04/16/linode_breach/ https://www.theregister.co.uk/2013/04/16/linode_breach/ Or again in 2015 when Linode's poor security lead to PagerDuty being hacked: https://www.securityweek.com/how-attackers-likely-bypassed-linodes-two-factor-authentication-hack-pagerduty https://www.securityweek.com/how-attackers-likely-bypassed-l... You can find some more discussion in previous HN threads about it: https://news.ycombinator.com/item?id=11136399 https://news.ycombinator.com/item?id=11136399
  - ryanlol 8y ago
    Yeah, others provided good links. I think your comment somewhat proves my point :) Nobody remembers this stuff!
nostrademons 8y ago
Or alternatively, if you're dead anyway. Many of the worst - and most strategic - examples of moving fast and breaking things come from situations where the feature needs to be written or the product launched, or else the company goes out of business. You need to hit the milestone to close the next round of funding, or you're dead. You need to add a feature to sign a customer, or you're dead. You need to launch - at all - or you're dead. It's rational for the milestone/feature/launch to be buggy if not doing it at all means there is no company. Remember that all startups start out "default dead", and getting to the point where they're "default alive" and killing the company is worse than the status quo requires a fair amount of effort and a lot of early decisions that may or may not be reversable.

by dsr_ 8y ago

credit_guy 8y ago

If breaking stuff causes a pedestrian to be killed because "they came from the shadows"[1], that appears to be perfectly survivable... /s [1] https://en.wikipedia.org/wiki/Death_of_Elaine_Herzberg https://en.wikipedia.org/wiki/Death_of_Elaine_Herzberg

ttty 8y ago
Cars killed people... they didn't stop creating cars. Nor would be a good idea.
- abraham_lincoln 8y ago
  No, we got jaywalking laws and our rights reduced.
  - anticensor 8y ago
    We got safer cars at the same time. It is a double-edged sword.
- TheCoelacanth 8y ago
  Killing one person per 1 million miles driven like Uber's self-driving cars have done is drastically different from killing one person per 80 million miles like human drivers do.
- drngdds 8y ago
  It would be a great idea if there was a good alternative in low population density areas.

dralley 8y ago

e.g. Knight Capital. Let's not also forget the consequences of "breaking stuff" when applied to industrial process control, avionics, car "autopilot" (cough) etc. "Survivable" is sometimes a literal term.

ryanlol 8y ago

>If breaking stuff causes your company to have a terrible reputation for privacy, security, or competency, that might not be survivable. I reckon this stuff is particularly survivable. You can be like Linode, fuck up everything over and over again and still do just fine.

osrec 8y ago
Or Facebook...
coldtea 8y ago
>You can be like Linode, fuck up everything over and over again and still do just fine. Source? Because every time I've read about Linode they had glowing user reviews, including on HN...
- detaro 8y ago
  a starting point (surprisingly high result on Google while looking for a list) https://news.ycombinator.com/item?id=10845278 https://news.ycombinator.com/item?id=10845278
- freehunter 8y ago
  It blows my mind that people still give glowing reviews to Linode when in 2013 hackers breached their systems and stole credit card numbers and customers reported their credit cards were being used maliciously and Linode got away with saying "nuh uh": https://www.theregister.co.uk/2013/04/16/linode_breach/ https://www.theregister.co.uk/2013/04/16/linode_breach/ Or again in 2015 when Linode's poor security lead to PagerDuty being hacked: https://www.securityweek.com/how-attackers-likely-bypassed-linodes-two-factor-authentication-hack-pagerduty https://www.securityweek.com/how-attackers-likely-bypassed-l... You can find some more discussion in previous HN threads about it: https://news.ycombinator.com/item?id=11136399 https://news.ycombinator.com/item?id=11136399
- ryanlol 8y ago
  Yeah, others provided good links. I think your comment somewhat proves my point :) Nobody remembers this stuff!

nostrademons 8y ago

Or alternatively, if you're dead anyway. Many of the worst - and most strategic - examples of moving fast and breaking things come from situations where the feature needs to be written or the product launched, or else the company goes out of business. You need to hit the milestone to close the next round of funding, or you're dead. You need to add a feature to sign a customer, or you're dead. You need to launch - at all - or you're dead. It's rational for the milestone/feature/launch to be buggy if not doing it at all means there is no company. Remember that all startups start out "default dead", and getting to the point where they're "default alive" and killing the company is worse than the status quo requires a fair amount of effort and a lot of early decisions that may or may not be reversable.