Hacker News Reader

Comment by ykm | Hacker News Reader

Halt production? It seems these soylent formulations are versioned. So can't they revert back to an older version till they fix the issue? Or its not as simple as this?

View on HN

Is it the version or the process? This isn't software, you have a lot of mechanisms & parties in play.

The biggest thing at risk here is reputation. Let me draw a comparison: how much worse would it have been for Samsung to just halt production of the Galaxy S7 then to race forward and produce a new, still-exploding version?

rollback both to a working combination?

in your analogy, it would be equivalent to Samsung putting its previous phone back into production.

Could be; in both situations there's enough external impact on the entire production process that reverting everything back to normal* still means people get sick.

* even normal could be in doubt. it's entirely possible this happened before and there weren't enough reports to warrant any concern.

>how much worse would it have been for Samsung to just halt production of the Galaxy S7 then to race forward and produce a new, still-exploding version?

Isn't that exactly what happened and it was a huge disaster.

they don't know what's causing it. If you just revert the formula, that doesn't mean the ingredient causing the problem has been removed. Contaminated soy protein is contaminated soy protein (for ex). Changing the amount of contaminated soy protein in your drink isn't going to fix anything.

The solution is obviously to create $n$ groups for each newly added ingredient in the previous version, randomly assign customers to a group and ship them a version with one ingredient added, then use R's prop.test to analyze the failure rate of the formulations. /s

I am one of the people who got violently ill from a bar. I ate numerous bars and boxes of 1.6. I had no issues, I usually eat 2.0 every day.

One single bar made me ill, I ate numerous bars before getting ill.

Reply on news.ycombinator.com