"We found that skilled forecasters who had access to machine-generated forecasts out-performed those who only viewed historical data. We also demonstrated the inclusion of machine-generated forecasts in our aggregation algorithms improved performance, both in terms of accuracy and scalability."
Either way, things went sour with the forecasts of recent major geopolitical events. Nevertheless, it is interesting that the best performing models were good old ARIMA and a state-space model.
"Through technology we are building connections between reality and the digital world. People and objects can be created to exist in parallel with a digital version of themselves in a digital world -- a world that mimics the real world. There is a growing impact of technology and digital connectivity on how we live, socialise, work, produce, and govern. AI is an emerging field in education studies, both in the EU and internationally. Several studies have highlighted the benefits of AI for inclusion -- pre-dominantly in terms of the added benefits associated with differentiated and 'hyper-personalised' learning. At the same time, massive digitalisation and data collection could cause (and has done so) safeguarding concerns and risks to individual privacy, while catering to the monetary interests of tech companies."
A good thing that they acknowledge ed-tech's pitches, but no critical reflection in the whole report about this so-called hyper-personalization. And no mention of smartphones.
ICANN charges a $185,000 evaluation fee to register a new gTLD, and they are under no obligation to approve it. They also expect you to actually run some infrastructure for your new gTLD, which has ongoing cost, and they will not approve your application if you don't have a plan for that infrastructure. There's also an ongoing fee of $6,250 per quarter per gTLD.
That initial fee is quite the racket. Quarterly's not as bad, but looking at some recent TLDs that have gone inactive you can tell those companies are desperate to save the cash:
Paramount let .cbs and .showtime expire in October, and Gap let .bananarepublic
and .oldnavy expire this month.
I'd guess more of them see that after ten years, Twitter still doesn't make a link when they post something like 'great content on whatever.cbs'. They have $6250 but run out of patience.
s/charges/charged/. The last time you could apply for a gTLD was in 2012.
> The next application submission period for new gTLDs is expected to open in Q2 2026.
> […]
> While the application fee has not been determined, it will be set on a cost-recovery basis. The fee will ensure that the next round of the New gTLD Program is fully funded and does not require funds from ICANN's operating budget. As a point of reference, the application fee for the 2012 round of the New gTLD Program was US$185,000.
In case of multiple applicants, there is an auction that can happen as well. In reality, various applicants settled among themselves so there weren't many.
1/3rd of the amount raised from the fee (60k USD or so) was earmarked for a legal fees. Some of the money is being spent on a ICANN Grant Program: https://www.icann.org/grant-program-en
A hot potato here, I am quite sure, and hence my take would be: some crazy things but some good, legitimate criticism too; yet I don't disclose which is which.
Either way, things went sour with the forecasts of recent major geopolitical events. Nevertheless, it is interesting that the best performing models were good old ARIMA and a state-space model.