We spoke with Tesla Senior Software Engineer Cyrus Yari about his views on software engineering, influenced by his own story of navigating the ruthless floors of investment banking and learning hard lessons as a cold calling salesperson in his teens.
Born in the heart of the old Persian Empire but made in England, Cyrus credits his mother with his astute decision-making ability.
She was a dressmaker in Iran before sensing the Khomeini regime would starve the country of opportunity and convinced her family that they must move to England.
“I spoke zero English when I landed,” he says. But within five months he was already winning spelling bees. “I'm pretty go go go. I got that from my mom.”
Rapid progression through practice-based immersive learning has served Cyrus well ever since. After graduating from Codesmith's program he was hired as a Senior Software Engineer at Tesla.
“People from Codesmith will succeed because they develop technical capacities at Codesmith AND already have a host of other soft skills”
Cyrus was the first in his family to go to university and studied Economics, Statistics, and Mathematics in London. But he quickly became disillusioned with the university model which he feels is an “outdated system.”
“I don't put university on a pedestal, if I started over again I wouldn't go. I think institutions like Codesmith are the next frontier.”
But it took some time for him to arrive at this frontier and software engineering as a career. University was followed by several years in investment banking, which he describes as “a horrific environment where your brain turns to cabbage.”
It was later, in 2018, that he started reading books by Charlie Munger and Nassim Taleb — “the two greatest thinkers of the last hundred years” — who redirected him towards investing his energy into skills with real leverage.
“It’s best to be self-sustaining in life and software engineering is one of the best things for that,” he says. “Technology is leverage. One can code something, let it run, and earn money while they sleep.”
But skeptical about the efficacy of another university degree, he instead looked towards a practice-based program to become a software engineer.
“The university model is all theoretical. But Codesmith’s model is practice, it’s built around learning how to learn and that's why it works”
“Taleb wrote that ‘for real people, if something works in theory, but not in practice, it doesn't work. For academics, if something works in practice, but not in theory, it doesn't exist.’”
Coming across Codesmith Cyrus was impressed by the high entry bar and distinctive focus on hard learning.
“I thought ‘Wow, I have to learn by myself and get to a good level — just to take a technical test to get admitted.’ It’s very unique in that it has a high bar of entry.”
Cyrus’ favorite way to learn is autodidactism, so Codesmith’s pedagogy resonated nicely.
“The university model is all theoretical. But Codesmith’s model is practice, it’s built around learning how to learn and that's why it works.”
However, just before he joined Codesmith, his plan was still to follow up the immersive with a Master’s degree in computer science, “I got an offer from a couple of places, one was Georgia Tech.”
But after tapping into the “Silicon Valley minds at the frontier of tech” hiring he realized it wasn’t in his interests to study computer science.
“They're all saying the best hires are those who just learnt through practice, out of passion.”
His choice to omit a computer science program was quickly vindicated as he was hired straight out of Codesmith by Tesla as a Senior Software Engineer. There he was able to see how university graduates struggled in the non-theoretical, fast-paced environment that is a major tech company.
“My friends working in big tech companies, and we at Tesla, hire interns (or did previously anyway) from university CS degrees, and they often can't cope, because what they’ve learned for four years is just largely theoretical.
This explains why the number one major pre-Codesmith for most residents is Computer Science — many from prestigious programs like CalTech, arguably the number one in the USA — with Codesmith providing what the university theory-based model lacks in practice.
“Most days in a big tech company you're assigned tickets, like ‘tackle this bug’ or ‘create this new feature.’ And how do you do that? By being able to code. Not theory. You can learn the theory that does matter online.”
Codesmith’s CEO and Co-Founder, Will Sentance, “has been to the Oxfords and Harvards, and I've been inside the Tesla's of the world. We’re both now against exclusive credentialism and believe in autodidactism, practice not theory.”
Many tech leaders whose journey began at a time before alternative programs were available struggle to understand their merit. Although Cyrus acknowledges many of these seasoned industry engineers from a computer science background still swear by studying the subject in university as the best route into tech, he playfully quotes Upton Sinclair. “It is difficult to get a man to understand something when his salary depends on his not understanding.”
“Aspiring engineers are in for a much harsher surprise by not doing something like Codesmith than by doing it. There's the hard way and then there's the harder way. The hard way is to do Codesmith.”
Cyrus’ technical capabilities alongside his soft skills like writing and public speaking has seen him move around Tesla to where the company needs him most, in a fraught year for the car maker.
He began working in Tesla’s Berlin GigaFactory. “When I joined Tesla it was awesome. I worked on a lot of zero to one products. They’d say to me "we need to build this, it probably doesn't exist, build it from scratch."
“For example, new software to catch manufacturing errors on the plant liner, or a 5G GPS tracker to find the most efficient route around the factory for the robots. But as well as coding, I also moved into doing a lot of non technical stuff like presenting to management, product management, UX and UI design.”
It’s leveraging these soft skills that has reiterated to Cyrus why coding ability alone isn’t enough to protect software engineers from a fluctuating tech market.
“At Tesla much of my job is human contact, you still need human beings. We are social creatures who crumble if we operate with no human element, as robots do. AI is a data robot, it doesn't think in complex ways of second order or third order effects.”
He explains that while the most gifted software engineers he works with at Tesla have impressive coding skills, they can be “challenging to work with, because they often lack the necessary people skills.”
This becomes most evident during Tesla’s weekly engineering review meetings, where new frameworks and tools for getting jobs completed are discussed.
“Rather than have a real conversation with the people in the room, they want to document and systemise everything. But the real world doesn't care about the coolest new framework. They care about what works.
“My brilliant manager at Tesla, a true practitioner, says ‘the customer doesn't care what type of paint brush you use to paint the house. They just care about the finished paint.’”
It forms part of his thinking on why Codesmith graduates are a boon to the tech industry and won’t find themselves replaced by AI.
“Most people at Codesmith are incredible. The type I would hire if I had a company, because they’re not one dimensional, like most of the engineers you find in these big companies who have been coding since their teenage years.
“They’ve had another career before tech. They've been a teacher, or an accountant, and have honed a lot of other skills.”
He acknowledges that in ten years AI could be doing miraculous things. “Even today, Devin, an autonomous software engineer, can take jobs from Upwork and pass technical interviews.
“But it’s still only 12% as capable as a fully fledged human software engineer. Ten years from now, it will be insane. Which is why software engineers, rather than just code, need to be like a Swiss Army Knife.”
“This is why people from Codesmith will succeed, because they develop technical capacities at Codesmith and already have a host of other soft skills.”
“Being able to only code is a danger in itself. Being able to code plus other skills is outstanding”
Cyrus sees the benefits and threats AI presents to software engineers today, and the type of engineer one is will determine whether their job is at risk from the new AI and ML tools available.
“The future of software engineering is AI doing the monotonous jobs for you. In that case, an engineer's job will become akin to a technical product manager’s role” — a capable product manager who also possesses a Codesmith grad’s level of engineering ability — having a bird's-eye view, architectural understanding of everything to ensure the pieces are connected properly.
However, he acknowledges that many different types of engineers make up the ecosystem, and some will be more at risk of losing out to AI than others.
The top 1% of engineers, the coding savant types, will always have a place due to their sheer technical skill, although, as Cyrus notes, they can sometimes lack the soft skills many companies desire in a modern software engineer.
The engineer who is a capable coder but whose only skill is coding, and doesn’t have transferable non-technical skills from previous life experiences and jobs, Cyrus says are at risk of losing out to new tools because AI will soon be able to do their work.
"Codesmith will thrive, as you learn the necessary real world technical skills at a fraction of the price of an Ivy League and don't go $200,000 in debt”
“Being able to only code is a danger in itself. Being able to code plus other skills is outstanding. It makes you antifragile - meaning you gain from disorder. When AI shakes up the market, you will gain as a generalist, a swiss-army knife, not a specialist who can only code and nothing else, the case with most engineers in big companies.”
The generalist, or “Swiss Army Knife-style” engineer, whose technical capabilities could perhaps be replaced by AI in the future, has a raft of soft skills alongside their coding ability, which AI will never replace.
This is the modern generalist engineer, for which the terms are still evolving, with a host of ancillary skills that Cyrus says Codesmith is adept at producing.
“These engineers would still need technical understanding, which is why Codesmith will thrive, as you learn the necessary real world technical skills at a fraction of the price of an Ivy League and don't go $200,000 in debt.”