How AI Algorithms Now Get Generated by Natural Selection

In this breathtaking post ‘Google Engineers ‘Mutate’ AI to Make It Evolve Systems Faster Than We Can Code Them‘, latest developments of AI algorithm generation is described.

It does not yet look like it really works for advanced algorithms, but there are possibilities that very quickly algorithms will evolve that will produce novel solutions to certain simple problems such as image recognition. This is quite an exciting – and troubling – development. It was due to arrive though with the development of ‘genetic algorithms’ that simulate natural selection.

Of course the novelty is now to apply it to AI algorithms which are by themselves heavier and more cumbersome to handle. Still it gives quite an interesting perspective of what we can expect in the short and medium term. Scary!

Share

How to Ensure Data Is Not a New Toxic Waste

There are quite a number of discussions about the ambiguous status of data in the new Collaborative Age. On one side it is celebrated as the new oil (refer to our post How Data Really is the New Oil, and Better); on the other side some argue that it is rather a toxic waste as in this interesting column ‘Data – the new oil, or potential for a toxic oil spill?‘

The point of the article is linked to data security and the harm that can be done through data theft and possible advanced recombination with other data sources that would also have been stolen. With zillions of data generated everyday, the argument is that one day or the other, sensitive data will leak and produce toxic effects on the wider data landscape and digital environment.

Specifically, the article mentions “Re-identification of anonymized data-sets [which] is a hot research topic for computer science today” and the fact that the breaches are additive in nature, progressively weakening privacy and sensitive data.

Of course, unclean data (refer to our post on data hygiene) is also another issue of toxic waste that may influence the wider data ecosystem if it is used as a basis for AI algorithm teaching or other reference applications.

The large amounts of data available today are a great source of value and at the same time are fraught with risks – as any new technology. Which will win first? My optimistic self is rather confident that the benefits will outweigh the risks, but that does not detract from the need to reinforce security and privacy.

Let’s make sure data is the source of value and not a toxic waste.

Share

How Data Hygiene Becomes Essential

In an AI conference lately I was struck by the mention of new jobs such as data hygienist and AI trainer. I did not realize how important data hygiene was – up to becoming a new profession!

Data hygiene is in reality quiet critical to AI development. Poor data hygiene is certain to create all sort of issues and false positive, and to lengthen dramatically the time it would take for an AI algorithm to learn its part.

Data hygiene is actually hard work because of the sheer size of the data bases to clean up, and the need to distinguish between rubbish and actual legitimate data points. It requires specific tools and particular attention, not to mention time. Hence it is a significant investment, but is found to be quite worthwhile apparently compared to the benefits.

Before we did not care so much about the quality of data in our databases – although there is still this old adage about garbage in, garbage out. Now we need a much higher quality level and apparently it is quite a challenge to achieve it.

Welcome to the world of data hygiene and data hygienists!

Share

How We Should Worry About Who We Have Helped Become Better People

Clayton Christensen, the innovation scholar that wrote so many books about innovation, passed away recently. He wrote “Don’t worry about the level of individual prominence you have achieved; worry about the individuals you have helped become better people.”

This quote if of course inspiring, and coming from someone with the life experience of Clayton Christensen, quite interesting too – since he had certainly reached a very high level of prominence.

This probably explains his drive to be a university professor while he could have had a very successful career in consulting and private ventures.

Still its reminds us that what people will remember is how much we helped them become better people. How can we work to achieve this better?

Share

How Easy It Is to Fool Artificial Intelligence

I love this funny post ‘Hackers stuck a 2-inch strip of tape on a 35mph speed sign and successfully tricked 2 Teslas into accelerating to 85mph’. The point here is not really about Tesla reliability, but how easy it is still to trick Artificial Intelligence recognition tools.

In this particularly funny example the researchers just changed slightly the speed limit sign and it was enough to trick the sign recognition algorithm that watches the road and determines what is the acceptable speed (see the image). This type of system is increasingly prevalent in cars generally just to update the actually applicable speed limit that is provided as a guidance to the driver.

What is really impressive here is obviously how easy it seems to fool an Artificial Intelligence-based recognition software. If that’s the case for something so obvious and mundane, then what are the consequences for more complex applications like face recognition? Are they also as easy to fool?

Artificial Intelligence does not seem to be quite completely robust yet. Some progress is still needed!

Share

How Irrelevant Industrial Age Approaches Are Still Prevalent

I read recently this account ‘Exclusive: Barclays installs Big Brother-style spyware on employees’ computers‘ of how a bank had installed a productivity measuring tool on its employees computer, that issues warning when people pause doing stuff on their computer. And indeed after a quick search I realized there seem to be quite an offer of “productivity monitoring tools” on the market.

This is an impressive application of the Industrial Age mindset as we move into Collaborative Age. Monitoring my computer activity would have absolutely no meaning as to my productivity: my work is about creativity, facilitating, getting people to work together. How can you expect to measure that based on my active interaction with my computer?

The article does not detail what were the specific tasks of the targeted employees, but in most modern organisations people don’t spend their entire day in front of the screen just repeatedly doing tasks that can be measured for actual productivity. Only some specific administrative departments could possibly be considered for that to be relevant.

In any case installing some software is a serious breach of confidence with regard to the employees and says a lot about the workplace culture that must be prevalent there.

In the Collaborative Age, productivity measurement must be more comprehensive than just interaction with a computer; and in case, trust will ever be a more essential characteristic of healthy workplaces.

Share

How Difficult It Becomes to Create Virtual Legends for Covert Operations

We already know ‘How Easy Modern Technology Makes It For Spies‘, it seems on the other hand that it makes it more complicated on some other aspects, like creating legends for secret agents. As explained in this interesting article ‘There Is No ‘Going Dark:’ Always-On Surveillance Posing Risks To US Covert Operations‘

In this fascinating account, US government seems to be quite worried about being able to maintain secrecy around its operations. Available data makes it relatively easy to know the background of individuals and what their real occupation is likely to be. Agents can easily be under constant surveillance as soon as they are in a foreign country. Biometric verification makes it much more difficult to change identities and travel under a different name.

“This is how spy craft works now. Everything is online, digitized, and likely to be accessed by agents of enemy states. There’s no flying under the digital radar. And if it’s true for government employees, it’s doubly true for US citizens who don’t have the ability to alter/remove collected data or a network of security professionals doing whatever they can to protect them (and their data) from outsiders.”

Even better, some recent reports show how critical it is not allow soldiers to carry their hand phone in battle, making it too easy to geo-localise them or to identify the source of their phone signal.

Our privacy is gone, and one consequence is that it is much more difficult for covert operations to be setup. It becomes increasingly difficult to make believe we are someone else than what really are.

Share

How Long Term Work Motivation is Related to Alignment with Purpose

Following up from the previous post ‘How Sexy Startups Can Also Sometimes Be Toxic Workplaces‘, one important aspect is that poor alignment with purpose generally don’t make it a sustainable venture. Toxic workplace cultures can’t be sustainable. Because if we want to remain individually motivated, our work needs to align with some purpose – and not just external motivators like compensation. This is developed quite well in Steve Pavlina’s post ‘Numbers vs Alignment‘

I find the point well written “If the numbers in your work (like sales and profits) matter more than the alignment of your work (like fulfillment, purpose, and appreciation), then even if you succeed on those terms, you may end up with bigger numbers but with lower alignment, which can strangle your motivation.”

And the corollary “People so often underestimate how much motivation matters – and especially how sensitive it is to alignment. It’s so easy to make misaligned decisions that eventually drag down motivation and lead to a place of stagnation, where it’s possible to be stuck for years.”

How aligned is your work with your purpose and what fulfills you? Is your motivation thus sustainable and not just temporary?

Share

How Sexy Startups Can Also Sometimes Be Toxic Workplaces

Startups are trendy and many young people dream working there in all the excitement of creating something that will change the world. Still some startups are also incredibly toxic places to work, as reminds us this Gapingvoid post ‘Beware Supersexy‘.

The point is particularly important to make because startups are by essence, stressful places to be where a substantial commitment is expected from employees, and change is prevalent on a daily basis as the venture grows, pivots and struggles. Like in projects, the pace of action is quick and some employees can sometimes feel overwhelmed.

Some examples have recently come to light of cultures of mental and even sometimes sexual harassment, and more benignly of certain toxic work cultures in some startups even some that were very much under the public eye and heavily funded.

It is not rosy everyday in startups, and the strong will that is needed from founders do not always translate in a nice way to work. At the same time I do know a number of startups that have developed very nice ways to work together and where employees are incredibly happy to the part of the adventure.

It’s just a fact that startups, like other organisations, can sometimes be toxic places to work, and that because of the specific pace it can be quite extreme. Startup are exciting, and also demanding. As always it is important to get proper insight about how it is to work there before committing to contribute!

Share

How Peak Income Comes Always At a Younger Age for Each Generation

One of the graphs presented in Lord David Willetts ‘Have the Boomers Pinched Their Children’s Futures?‘ (refer to previous post ‘How Demographics Can Only Explain Part of the Millennials’ Economic Situation‘) has struck a particular chord with me. In this graph we can see the revenue profile by generation according to age in the UK.

In the speech David Willetts uses the graph to show that for the first time younger generation’s income is lower than previous generation. However I find that the most interesting part of the graph is how it shows that peak income for each generation arrives at a younger age. For people born in the late 1940s just after WW2, it happened in their late 50s; for people born in the late 1960s, when they reach about 40. This all points to a ‘peak income’ sometimes in the decade 2000-2010 with income decreasing later for all.

This observation explains why those that arrive now on the employment market won’t find the same level of opportunities. It is also a question mark about current income levels – across all generations.

Therefore the interpretation of the graph is rather that there is a general loss of income in the last decade. It impacts the start of millennials’ careers as well as all other active generations. This is the key issue that needs to be tackled, rather than inter-generational transfer issues.

Share

How the Collaborative Age Requires A New Leadership Mindset

This MIT-Sloan paper ‘Leadership Mindsets for the New Economy‘ takes the perspective that the new economy requires a shift in leadership practices.

It starts with the excellent quote by Patty McCord, former chief talent officer, Netflix: “In today’s world, everyone has to adopt a leadership mindset. We have to think of ourselves as members of a leadership community“. This means that it is recognized that in the collaborative age, leadership capabilities need to be more widely spread inside organisations.

I find the rest of the paper a bit disappointing and too MBA like, with the identification of four key traits of leaders in the modern economy – producers, investors, connectors, and explorers. It does not go back to the question of how to make everyone in the organisation a leader – and how to make sure everyone plays the part he or she is the best about among those four traits. And that’s clearly the most important.

While this issue is recognized (“building a collective leadership capability is the strongest route to competitive advantage in today’s fast-paced world“), tomorrow’s determining leadership trait is indeed to allow the growth of leaders in all levels of the collaborative organisation. I’d rather see research exploring that direction.

Share

How Excellence is a Moral Decision

In this Gapingvoid post ‘Creating excellence is not a job. Creating excellence is a moral act‘, the point is made that “Excellence is not a law of physics. Excellence is a moral act. You create excellence by deciding to do so, nothing more“

This means that excellence is what you do when no-one is looking, and it is a personal commitment. It can even become one way to define onself like Horst Shulze co-founder of the Ritz-Carlton Group is quoted saying “And life becomes much more valuable. It becomes much more fulfilling. It becomes something where you’re using your time to define yourself, and the first one who will see it and will be happy about it is you, yourself.”

This also means that striving for excellence can’t just be imposed from above by a manager. It is a real leadership act and requires leaders to demonstrate their commitment too in everything they do.

Excellence is not a quick recipe and a buzz word. It is a moral decision and requires strong leadership to spread.

Share