The "Promises" Of Automation

About The Last Two Weeks...

Last weekend, I went to brunch with a couple of very dear friends and one of them, the incredible Dr Suze Kundu, was kind enough to introduce us to the Indian festival of Holi. As we were waiting to cover each other in coloured powder, Suze made a point of thanking us for sharing the celebration with her. I actually felt that gratitude should go the other way (and said so), that we were grateful to be invited to take part in it, and suggested that if it were up to me we’d have bank holidays for all religious and cultural celebrations. How better to understand others than to respectfully take part in what’s important to them?

The idea of a large number of bank holidays stayed with me and merged with another conversation we had that day, about A.I.s. And it had been gnawing at me ever since. More free time was one of the promises of automation. Doing the boring things you don’t want to do.

But so far, it seems that the A.I.s we are getting are not really focused on the boring parts. They are focusing on making art or explaining things or writing poems or even praying for us. I am not personally or financially affected: my poems are terrible and although brought up Catholic - so I can administer some sacraments in case of emergency - I feel a computer might offer better prayers than I do.

I do create art but I doubt it will interest those who train these algorithms, and the training is the problem. These algorithms are designed to mimic some talents but they are not providing something that the people truly need. They are shortcuts that take away work from artists, but they do not fulfill any real need for the workload of the average person.

But my job is safe for now given the screw-up Google's A.I. had about science:

Google’s Rival To ChatGPT Makes Embarrassing JWST Error That Wipes $100 Billion Off Shares

AI’s unrelenting takeover of all humanity has hit a minor hurdle.

IFLScienceTom Hale

For Google and Microsoft's boastful claims of the beginning of a new era of the internet, of computers, etc. I remain a bit skeptical. It's not that these tools can't be revolutionary;  I have seen positive things about using ChatGPT as a code-writing helper which is pretty cool, and Microsoft Copilot seems like the extremely smart successor of Siri, Alexa, etc. - Copilot seems to at least focus on streamlining workload, but even there the focus seems to be off. And don't get me started on the ethics.

Microsoft chief Satya Nadella introduced Copilot and said that "For the first time we have the access to AI that is as empowering as it is powerful. Of course, with this empowerment comes greater human responsibility. Just as an individual can be aided by AI, AI can be influenced positively or negatively by the person using it."

This line doesn't sit well with me. We need more than personal responsibility. We need corporate responsibility that the A.I. is ethical. And the fact that Microsoft fired its entire A.I. ethics team doesn't fill me with confidence.

I have not seen bold dreams or proposals about how this tech can help humanity. It seems to be all about "money and productivity but please don't ask us about ethics."

WHAT I LOVED WRITING THIS FORTNIGHT

Given how A.I.s are the hot topic of the moment, it was easy to pick relevant stories for the newsletter. It turns out that there was a lot of interesting stuff. But two stood out for different reasons.

First up, this story is about predicting neutron star collisions. These events are responsible for the creation of heavier elements such as gold and platinum. These mergers can be observed with gravitational wave observatories as well as traditional light ones. The proposed methods let scientists know that a collision is imminent. They will know if is it about to happen tens of seconds before it happens. It might not seem much, but it might lead to actually seeing such events live.

Can We Predict Incredible Stellar Collisions Before They Happen?

To study collisions between neutron stars, every second matters.

IFLScienceDr. Alfredo Carpineti

Since the last newsletter I have been at Virgo, the gravitational wave observatory in Italy (more on that at a later date) and so seeing this research come out, I definitely wanted to write about it. I love the different ways automation is used to speed up astronomical observations.

And talking about astronomical observations... A.I. might fake them. A Redditor has accused Samsung of cheating with the Galaxy phone's Space Zoom. The company claims that it's all about the camera system and the software stabilizing the photo, but this Reddit user has shared evidence that the software is using an algorithm to overlay a good picture of the Moon on the shaky ones you'd take.

Redditor Accuses Samsung Of Faking Its “Space Zoom” Moon Shots

The promos show incredibly quality pictures but there might be AI trickery at work.

IFLScienceDr. Alfredo Carpineti

Now, my question for you is - let's ignore the possible false advertisement. Does it matter that the image of the Moon is not exactly yours? It's not like the Moon changes... and personally, I think it would have been very pretty to have a clear full Moon next to the tower of Pisa in this photo!

Feel free to tweet at me or message me if you think it's a good idea or not to have a little helper for your social media photos.

Reading Reccomandations

Not exactly a recommendation as I am yet to read it, but a book just came out called Hidden In White Sight, discussing specific issues of how artificial intelligence can display human biases and prejudices - yes, technology can be racist if how we build it and train it to be so.

Hidden in White Sight | How AI Empowers and Deepens Systemic Racism |

Artificial Intelligence was meant to be the great social equalizer that helps promote fairness by removing human bias from the equation, but is this true? Given

Taylor & FrancisBy

Recipe of the Fortnight

Lime & Dark Chocolate Cookies

So in the last two weeks, I made this recipe on a suggestion from Craig (@Pokubakes) and it turned out pretty good. Although personally, I would add a lot more lime zest, they were pretty yummy and they didn't last long at all!

If you want to try and make the recipe, here is the one I followed.

Lime Dark Chocolate Chunk Cookies - Cook Nourish Bliss

Perfectly soft and chewy lime dark chocolate chunk cookies! These irresistible cookies are filled with lime zest, chocolate chunks and flaky sea salt!

Cook Nourish BlissAshley

From The Astroholic Explains

Unlike last time, I do have a relevant episode of The Astroholic Explains with A.I. - Pretty much there are many efforts to create machine learning algorithms that can do what I did in my Ph.D. but better. I don't think I am obsolete just yet, but maybe soon...

And here is lovely Clue!