Got asked in a PM interview: Build an AI Powered Killer Robot.
Was asked this question in a PM Interview at a unicorn Fintech based in Bangalore. I tried my best to answer but just trying to understand what PMs here would try to solve this?
Talking product sense with Ridhi
9 min AI interview5 questions
I'm not a PM but if I were to answer I would ask about the user base (just to show that I am user focused, I want to get to killing right away). Then I'll think of cheapest solutions. Bombs,guns etc are operationally expensive. So I'll go with poisoning. Easiest method will be gas.
As for the robots, putting AI in every robot is again expensive. So instead I would prefer a fleet of robots, controller by an AI. Decision making is centralised and it is easier to stop in case of any bugs.
But gas propagation is also hard to control, so we might want to use data analytics to figure out where and when the majority of our targets will be together.
Example, if they all are same company, are they of same educational institution etc and try to optimise it.
Each robot will keep count of the kills, misses etc and we will use it as feedback to improve our next versions. Monitoring and feedback is essential
(I have no idea what PM does, I just wanted to try making cheap killer robots)
Now that I think about it, gas is easily preventable with masks. So we need something that can go through masks. Maybe radiation, but then there's no way to control it (with current technology atleast).
Then what? Poisoning the water? But they can figure out filtration.
We can use supersonic sound waves to burst their eardrums maybe. But those machines can be expensive and operationally expensive to carry around as well.
Maybe we can use AI, to turn each targets digital assets to zero. They become broke, can't afford anything and are driven to kill themselves. This might not work on everyone. Low success rate.
Well well, everyone please tell your ideas as well
@SuperNova This is actually such a detailed answer that I will use this in parts
Prime Minister or product manager ?
🤣🤣🤣 Best!
@Momo_chutney I think Product Manager
Not a PM but adjacent function (PMM). Such questions are meant to throw you off, do the framework to answer. I’d first laugh and ask why who is the user etc while thinking ye kaisa chutiya hai
Then I’d:
Understand needs - population control? Crime control? Terminator in real life? :p
Prioritise what to build to meet end goal - most kills least spends
Solution - link how features will achieve what
Trade offs - AI evaluation can be flawed - add human reviews before killing
Summarise
@chillpanda This is actually not a bad question
@Jackietrader Didn’t say it was. But I don’t see why you’d make young IC candidates suffer such hypothetical questions.
This one is a bit nasty since if the candidate is a “good person” they’d be battling their conscience about AI killing humans before even attempting the solution.
A robot which kills when asked dumb questions
@Mercurial87 Bwahahahah
Naye bando ko hire hi na kare you mean 🙄😂
- Ideation
- creating user persona. Who is our target audience
- in this case mostly it will be states, may be as weapons, stealth or otherwise
- It could be mafia who are willing to pay
- This is crucial for requirement gathering, marketing and all other feedback.
- create use case
- Assume it is to be stealth,
- the the robot should be able to kill
- robot should take instructions to kill individuals
- it should be able to identify individual
- it should be able to kill with few methodology, by gun or using poison
- Killing should be stealth and shouldn’t be identifiable
- Assume it is to be stealth,
- creating user persona. Who is our target audience
- Validation
- create a sample robot which can kill dummies
- kill in various condition like still, moving,
- able to identify correct dummy among multiple similar dummies
- Able to register poison or kill with bullet
- Create a matrix of data which can be used to evaluate the product
- Discuss the possibility of creating all these features with engineering team.
- Check with finance team for budget and auditing
- Check with legal team on possible complication in future
- Prototyping
- create a small robot with features gathered in validation stages
- Create a performance matrix
- Create basic data for al possible feature with engineering team, to catalogue and to sell in the market.
- Discuss the pricing with internal team.
- Marketing
- take the features to create a GO TO Market strategy.
- Meet possible customers, understand their requirement, propose the solution
- Try to sell by educating.
- Development
- Further enhance the feature on performance matrix.
- Enhance based on feedback
- Add features
- Launch
- Launch full fledged product to larger audience with better marketing
- Create proper communications
- Package the product
- Manuals etc
- Improvements
- keep on improving
Please review
Nice answer but no details
