Profile cover photo
Profile photo
Joni S.
Joni's posts

Lots of attention is put to 'AI replacing humans', but the reality is current systems are enhancing knowledge work rather than replacing it. #algoritmitutkimus

For example, they support the work of doctors, journalists, and marketers, but you still need the people to make the final calls. The systems just empower individuals to do more. It won't be anytime soon that we have a strong artificial intelligence that would replace knowledge workers; rest assured.

Startups are often credited of having a "higher innovative capacity" than corporations. However, if you measure by patents - not the case. The same if measured by impact - very few startups actually make any difference. Amidst of hype, gotta remember it's still the corporations that do the major innovations and call the shots.

What are the ways for platforms like @Udemy to reduce transaction costs? Thinking of project managers, liquidity monitoring, automatic evaluation of task descriptions, recommendation systems...

Fighting online polarization: finding common ground


A is of opinion X
B is of opinion Y

X and Y are of opposite extremes, therefore A and B are against one another

How to solve?

Find C = unifying factor for A and B

What is C?

It can be a person (e.g., common friend), hobby (e.g., football) or another interest (e.g., Hellsing anime).

Solution is to find communality to alleviate polarization.

Ideas for filtering out low quality participants from MTurk surveys (Anglim, 2013):

- item-level reaction time (too fast reaction --> remove)
- response patterns to negatively and positively worded items (too inconsistent --> remove)
- repeating items which should yield identical responses (too inconsistent --> remove)
- simple true-false questions (e.g., 2+2; Who is the president of the United States) (inability to answer --> remove)

Twitter does such a service to the society by providing an open, accessible source of information. Huge source of sentiment data, much more than Google and Facebook which are only semi-open. But the monetization is so bad, so I fear for their future. They should do like Reddit, and force ads around "topic communities", proxied through hashtags; and like Snapchat, by selling premium campaigns that can be both site-wide and topic-focused. In any case, the topic data is their key.

startups as lego bricks

break to pieces
build again, different model
break to pieces
change some pieces
build again, different model

While doing research for my dissertation, I realized there are multiple layers of failure factors associated with any given company (or, in reverse, success factors).

These are:

- generic business problems (e.g., cash-flow)
- individual-level problems (e.g., personal chemistry)
- company type problems (e.g., lack of funding for startups)
- business model problems (e.g., chicken-and-egg for platforms)

Only if you combine these multiple layers - or perspectives - can you understand why one business venture fails and another one succeeds.

There are no easy answers, although the human mind and various success stories would lead you to believe so!

Many unsuccessful CEOs are not "customer people", they are optimizers. They want to squeeze more value from a big machinery instead of growing the revenue base by focusing on customers. That often works, because somebody else has thought of customers before -- the corporation has inertia in that sense. But when the inertia stops, the optimizer will fail.

A major fallacy in "robots are taking our jobs" is assuming number of jobs is constant. The whole point of capitalistic economy is 'creative destruction' - from labor perspective it means new jobs are created to replace the old ones. Thus, the question is not "Are robots taking (current) jobs?" but "Are enough (new) jobs being created?". Policies need to be done so that money moves and people buy and offer services.
Wait while more posts are being loaded