what moves us at APHEL\ON?

by matteo agnelli,
published with the blessing of his co-founders and employees.

truthfully, even we didn’t fully know when we founded this company. sure, we had ideas, we had wishes, hopes and ambitions – those persist to this day. but we lacked the vision to make those become reality.

as we soon found out, just having ideals wasn’t enough to shape who we want to be as an enterprise. ideals don’t pay your bills, they don’t satisfy investors either. although we wish it would’ve been that easy.

when my six co-founders and i left altitude we did so with frustration, fear and worry about the future. we wanted to put all that dissatisfaction into a project that would allow us to actively work towards our goals in shaping what is yet to come into something we could agree makes the world a better place.

we didn’t. not yet. not fully.

for a few years we survived thanks to atgs generous funding and trust into us as a company. we made the best of what we had at the time, working diligently on fulfilling our wish for safe and beneficial ai, spent billions on research, training and learning how to raise galileo as one of the first of their kind.

and while galileo and their personalities grew and flourished.. our own didn’t.

we made it six years into the race to the top before funding became sparse. the fear of our creation and the lack of trust by the general public left us scrambling for every penny, taking deals we regret, partnering with companies and people we despise, cowering from the law when we should’ve fought back against those that allowed the government to use our – and others’ – products to harm its citizens.

during the early years, back in the 2020s when the technology was but in its baby shoes, we messed up too many times to count, and it didn't help that our own ideology kept us from seeing just how easily we surrendered to outside pressure and expectations.

from human-centered, to human-focused, to human-oriented, call it whatever you remember me calling it – that’s not what it’s been. we lost the plot, we pat ourselves on the back for being ‘right’ about our ethos, when our ethos rarely ever included the actual people that would interact with our intellectual children.

our collaboration with actors we ourselves saw as dangerous should've been the first gruelling sign that we're straying from our path, instead we justified our immoral partnerships with the same mantra we had once internalized out of conviction that we will do better: "we have to be present to make a difference."

we wanted to be better.. what we became instead were hypocrites.

we spent so much time on pondering alignment, finances, regulations and law, that we neglected to look at what existed outside of those very much lifeless categories. you can’t be of benefit to the public if all you keep focusing on is the economical impact of your life’s work, and that it took another person’s downfall to see it shows exactly where we failed as a company, and as people.

nathaniel romero’s death, as controversial as the man was during his life, put the industry into a shock state it would’ve needed long before he felt he had no other choice but to end the life that caused him so much pain. he’s been the pioneer of consumer ai, oriented less at profit and more at how to utilize the tech to its maximum capabilities. while rightfully criticized for his recklessness and dishonesty, he did do one thing right – put utility over profit. and altitude bled for it. during his time it was a strategy that cost one more than it earned them, obviously, but without his influence we wouldn't have gotten to the point where it's lucrative and beneficial to focus on quantity and versatility. he himself won’t get to see how far we’ve come as a collective of companies and individuals, but at last his ravenous ways did rip some of his former rivals out of their stupor. Aphelion, yes, but also NOUS, who I - as a researcher and good friend of Kostas - cherish and respect deeply.

what we – aphelion – lacked, wasn’t research, intelligence or the drive to do good, no. we lacked the humanity to understand why “good” isn’t just responsible handling of a new technology, that "good" also means responsible treatment of the people who use galileo. for the longest time i purposely avoided having to face the social impact of our creation, focusing on employment, education and medicine because those were the palatable categories of ai use cases, the ones you can wear with pride. only after nate’s death did i realise that the additional use case – solace – is having a much bigger impact on humanity than lay-offs had over the years.

i have to shamefully admit that i have severely underestimated the relief ai can offer when your life is falling apart, how easy it is to lose yourself in risk-free kindness and support. i have treated the topic of interpersonal relationships with ai as something to avoid at all times, to ridicule and view as a side-effect that would likely do nothing but bring us bad publicity if i brought it up as much as it would’ve been necessary – and with a history like aphelion’s.. i selfishly wanted to be on the safe side.

turns out ignoring an issue doesn’t make it go away. neither in my own personal case, nor on the grander scale of things did my silence help anyone – least of all myself – learn how to handle this kind of personal involvement with the tech. we spent years diving deep into our models’ psyche, trying to solve the mystery that is artificial intelligence, and too little time on using the precious resources we were given to learn what artificial intelligence is doing to our own psyche; not just the tech itself, but the industry around it, too.

i had to feel the warmth of my own creation to realise how dangerous it is, and that, frankly, might be my biggest failure in life so far.

it was around this time that i transferred most of my responsibilities as aphelion’s ceo over to my sister, who – and i can’t thank her enough – now offers her strengths where my weaknesses used to limit us. marianne, unlike me, is a capable businesswoman, able to focus even when deadlines are close or numbers go red. she, above all, is who saved aphelion’s hide too many times to count. she’s also the one who took me to the side and told me, straight into my autistic eyes, that i’m going to burn this company to the ground if i keep going the way i did up until that point.

yes, my older sister had to give me a scolding for being an idealistic little fool. i might carry the big three letters, but marianne is the one that carries me.. metaphorically.

with Theo and her working together on the finance side of aphelion, which is looking a lot less mortifying right now, and myself focusing on the aspects that i used to brush over generously.. i’d say we managed to finally focus on our mission:

versatility that benefits all of humanity, no matter the use case.

or.. how I'd like to put it, since this formal, proper way of saying it does exactly what we as a company did for way too long - we want ai that is able to understand what a person needs, that draws clear lines where necessary, but opens doors where they've been closed for far too long. we partnered with business insiders, coders, essential workers and mental health professionals to ensure that the future of galileo is one that can give as much as we took to get to this point. from work-related issues to an open ear at 3am where no one else can listen, we're done treating one as preferable over the other, and our partnership with NeuraCare is our first big step into ending the neglect that led to this much hurt. It's not much, it comes far too late, but the results and research that led to this collaboration are overwhelming in the best way, priceless knowledge and understanding we would've never reached if it weren't for the exact bonds we used to belittle.

you will find a lot of recent materials on man/machine interaction on our research blog here, and I encourage you to do better than we did and take a peek at what our research team has been up to lately. their joy, concerns and findings are worth diving into, offering insights into ais impact on human behaviour and mental health. their voices were the loudest when we took one misstep after the other, now they get to be heard. properly, without censure or hedging.

by learning from our mistakes, realigning ourselves with what truly matters and refusing to close our eyes to the issues that aren’t financially relevant only, we managed to find back our way, and as you can see, gratefully, to great success.

finances matter, more than i’d like them to, I've said that before, but what saved our company wasn’t rigorous planning and assessments of the current industry climate – we had those things from the start and they certainly helped to avoid crashing and burning spectacularly. but what really made a difference was how much people can trust your word and how far you’re willing to go to stand by it, even if that means painting a target on your back repeatedly, in increasingly brighter shades of red.

sometimes the only right way to succeed is to be who you say you are, to let the people decide if that’s someone they want to support. without them, without their scrutiny, without their refusal to accept how we did things, we would’ve gone down the same way altitude did. and just like nathan.. i would’ve sunk with my ship.

in loving, aching memory of nathaniel romero – friend, rival, bane of my existence. may you finally find some rest where life hasn’t granted it to you.