Falling between the cracks
Jon Fisher

Could a product have the capability of killing someone?

Three real life examples were given:

  1. Chernobyl
    A mixture of poor design and human behaviour led to the core in one of the nuclear reactors exploding. This was caused by an optimising violation where someone attempts to break the rules with the intent of achieving someone good. In this example, the engineers wanted the safety test to pass so they broke crucial safety rules to do so. In total 31 people died (if you believe the official statistics).
  2. Railway Safety
    Unfortunately, I did not write enough down to fully remember or understand this particular scenario but it involved someone working at a computer where they had to perform repetitive tasks. A chain of events led to the person at the computer making a mistake due to the repetitiveness of his work. The railway line became fully electrified while an engineer was doing maintenance work. Fortunately, no one died. Unfortunately, the engineer had to have both his hands amputated.
  3. Ethiopian Airlines Flight 302
    The cause of this plane crash is still under investigation, however it is generally believed to have been caused by a sensor recording the wrong flight angle. The computer decided to dip the plan to correct this angle. The pilot noticed this and attempted to stop the plane from dipping. The pilot and the computer were fighting each other – the computer won the fight and 157 people died.
The Swiss cheese model was mentioned as a way to show that there will always be several holes in the design. Accidents can happen when those holes are perfectly aligned.

When designing a product, the desired outcome is to deliver value to the customer. Is there an obsession with value? Are we even aware of the potential risks and pains involved when delivering that value?

The human will try and do things the tech team believed they shouldn’t and won’t do. Humans are unpredictable – they they probably do the unexpected. No matter how many levels of defence, there will always exist that perfect chain of events that can result in catastrophe.

I have one observation with the examples mentioned above. We have 2 situations where a human was trying to fight the system. In one case, the human won resulting in Chernobyl. In another, the system won resulting in the Ethiopia plane crash. Do we design to allow a human to take over when required when the computer has gotten it wrong? Or, do we design to prevent a human taking over so they don’t do something stupid?

Web 3.0: How blockchain will change the way we interact with one another
Georgia Rakusen

Thew world is full centralised systems who control everything we do. What is the problem with centralisation? Everything is controlled by a central organisation who have all the power. All information is controlled by the central power, which can create questionable integrity.

Centralized vs decentralized vs distributed processing
Centralized vs decentralized vs distributed processing

Blockchains allow information to be stored across a network of computers. Because the information is not stored at a central location, it is not owned by a single person or company. Multiple people are encouraged to cooperate to verify the information and transactions are valid. Since the information is stored and checked by multiple sources, the overall system has better integrity.

A few examples were given where such a system has been beneficial.

CIvil – The journalism industry is reliant on ad revenue which influences content. As a result, we have no idea what information is correct or not. A decentralised system of co-ownership and participation can help build a more integral industry. Members have to follow a code of conduct and can be voted out if that code is broken.

Openlaw – Normally, legal contracts are controlled by a lawyer. This can make any legal process slow and cumbersome. Instead, legal agreements are created and signed on a block chain. Without any central lawyer, it is easier to raise disputes. Ultimately, all parties involved have to agree.

UPort – an open identity system where personal information can be easily transferred to new platforms. The user has better control over what information they want to share, and what information they want to hide.

I found this great video explains what blockchain is better than I do.

I hope you enjoy reading my summaries. I find its a great way to review my notes and record my own interpretation of the talk. Next post will be about the talks by Gavin Strange and Hilary Brownlie.