Overview
Saoud Khalifah is a co-founder of Ciphero [1], a platform based in New York that focuses on AI verification and governance for enterprise AI interactions [1]. The company operates within the New York artificial intelligence sector, addressing enterprise needs related to the verification and governance of AI systems [1].
Career history
Insights & ideas
The through-line
Saoud Khalifah's account of his own work returns again and again to obsession as the actual engine, and to usefulness rather than money as the reason for building. He is explicit that Fakespot was never a wealth-compounding exercise: "with fake spot I was not doing it just to make money I was doing it to build a product that would be used by many people to bring some kind of benefit to their everyday life and that's how I approached it" [1]. He contrasts himself directly with founders who operate strategically across ventures, saying of Elon Musk that "every company he did he compounded his wealth to put it into his next company," while "I was more interested in building stuff that was useful to many people that was my I kind of now realize that that's my meaning when I'm doing this type of uh software" [1].
Running alongside that is a second constant: a refusal of comfort. The through-line from teenage reverse engineering in Kuwait to leaving a Goldman Sachs job is a repeated decision to step out of a settled pattern into an unknown one, and he treats that instinct as the thing that separates people who build companies from people who do not [1].
On obsession as the real mechanism
He describes the founding period of Fakespot in terms of hours rather than strategy. Working as a quant at Goldman Sachs with a two-hour commute each way from New Jersey to Wall Street, taking the ferry, he would reach his apartment at 8:30 p.m. and "work on fake spot till 3: a.m." His own summary of that stretch: "that was my life and you know I had nothing else going on Obsession just an obsession with this" [1]. The same pattern predates the company. He was coding from age 11, starting with a Dragon Ball Z card game in Visual Basic 6.0 built for himself and his brother, and describes a standing habit of trying to program anything in his life that could be automated [1].
That habit is what he credits for Fakespot's existence at all. Fakespot began as one of those side projects, launched for himself, not as a business: "when I launched fake spot it was just for myself first uh I wanted something to help me when I'm shopping online and then you know as I was building it I had a lot of passion for it I didn't really care about building a company" [1].
On risk, uncertainty and leaving your environment
Asked why travelling young correlates with becoming a founder, he reads it as a proxy for risk appetite: "people that don't like to take risks uh will remain in their environment uh probably their whole life and find comfort and solace in their own environment that they're very acquainted with uh folks that like to take risk will try to venture outside" [1]. He extends this in two directions. Culturally, he argues the disposition is inherited: Arabs "were Nomads for a very very long time," moving between regions without knowing whether the next one was safe, in search of trade or support for the family, "so it's in our genes believe it or not" [1]. Demographically, he points to America as having a higher ratio of such people because "everyone kind of immigrated to America and they had to leave their comfort zones," and says such moves correlate with a higher inclination of certain brain chemicals [1].
He does not treat this as pure determinism. Founders, in his account, "like to assess risk and then take gambles and Bets," and the trait is "an Arsenal attribute of uh in your character that you you can build," even if some argue certain attributes must be innate and only honed with maturity [1]. He also names the pressure that develops it fastest: in a startup hot seat, "the character develops uh like a rocket ship," because you are simultaneously building the product, the team, the culture and the business across different tracks, and that develops traits he considers critical to startup life [1]. He notes he was among the very few in his cohort of Kuwaiti, Saudi, Qatari and Emirati peers in Oregon and New Jersey who did not return home, and frames his own decision plainly: "I was in my 20s and I'm like hey there's I have nothing to lose let's go all in" [1].
On why he left Goldman Sachs
His critique of Goldman is not that the work was bad but that it was legible. "It was a little bit too comfortable for me," he says, describing being a cog in an established machine over a hundred years old where "everything there was a pattern to everything there uh everything how you move up the ladder how you talk to people the culture there was a pattern and everything was how you dress," down to ironing a suit every day [1]. He notes he was the only Kuwaiti on Wall Street at Goldman at that time, and was curious about the cultural differences between the New York and Riyadh offices that some teams worked across [1]. The pattern itself is what pushed him out and into starting his own thing [1].
On hacking, vulnerabilities and how software actually works
His formative technical interest was breaking things apart rather than assembling them. As a teenager he moved from game-making to security, hacking and reverse engineering, drawn specifically to assembly and machine code: "I just loved breaking uh uh software down to its core elements and this was the machine code essentially and you could find uh holes within that software" [1]. His brother's reaction to reading over his shoulder was that it looked like an alien language [1]. He singles out buffer overflows as one of the biggest and still most prevalent categories, vulnerabilities where a payload such as a file lets you take over someone's machine, and notes these are what intelligence and three-letter agencies use to infiltrate their counterparts across regions [1].
He found bugs in Microsoft and Adobe software while still at school and later while at university, and was paid through Zero Day Initiative, one of the first bug bounty programmes in the world, buying his first car with the proceeds of a Microsoft PowerPoint vulnerability that allowed a shared file to exploit a buffer overflow and take over a computer [1]. He tells the Microsoft internship rejection as a story about institutional culture rather than about himself: under Ballmer, "if you kind of damage our reputation we do not want to work with you," where he had expected they would want him inside helping fix things, and he observes that the attitude has since changed [1]. He also notes that Kuwait had no real industry for programmers and security when he was there, and takes evident pleasure in the fact that one now exists, with Kuwait University people forming a security group and researchers who attend Black Hat emailing him about the vulnerabilities he found in 2008 to 2010 [1].
On the review problem and building Fakespot
The origin was a specific betrayal of trust. In his last year of college he bought an endurance supplement on Amazon with hundreds of five-star reviews, "the yellow stars were glowing in my face," and received something that looked garage-made, with tape falling off and pills containing what looked like sawdust, which he did not dare ingest [1]. Going back to the ratings, he found many were bots and many praised the company rather than the product, and concluded: "this is uh like my moment where I realized I can't trust reviews anymore on the internet" [1].
The product design followed directly from watching his own manual process. After spending hours opening a new tab for every reviewer profile, he asked "wait a second what I'm doing right now I can automate uh with with AI like emulate some parts of my brain with a couple AI models one that will look at language another one that will look at the purchasing behaviors another one that will look at the company" [1]. The first version was deliberately minimal: a text box and a button, paste in a product page, and within thirty seconds the models retrieve the reviews, the product and the company and return a grade on whether the reviews can be trusted [1].
On naming things
He treats naming as a first step rather than a later one. "I like to have a name before a project starts uh it's just very similar to naming your child," he says, citing a claimed correlation that a child given a strong name is more predisposed to success, which he characterises as a psychological effect [1]. Acting on that, he pursued the taken fakespot.com domain around 2015 or 2016, tracked the owner to a kid in the American Midwest who ran a Counter-Strike clan called fakespot, and bought it through a broker for a little over a hundred dollars [1].
On environments that shape technical people
He is attentive to the people and institutions that formed him and treats these as partly luck. His grandfather in Poland was a programmer who gave him his books and got the family a computer, on which he played Doom on MS DOS, which is how the interest "kind of snowballed" [1]. At Monmouth University in New Jersey, many professors were ex-AT&T Bell Labs, having joined before the breakup of the monopoly and moved to the university afterwards, each specialising in a different area such as neural networks, UI or next-generation user experiences. "I kind of locked out to be honest with you I didn't know about the background of all these folks at this time" [1]. He draws a live parallel between that antitrust breakup and what may now happen to Google and other companies, while stopping short of predicting it: "time will tell" [1].
On believing in a future before it arrives
His conviction about e-commerce ran well ahead of its adoption where he lived. As a teenager in Kuwait around 2007 and 2008 he ordered books from Amazon through a New York City PO box forwarding service, "that's how much I believed in this future in Kuwait you know at that time I remember no one knew what was like what that was" [1]. He connects this to how he thinks great builders operate: they "looked at a future that was their own version of a future and they put everything into it" [1].
Takeaways
- Build for usefulness first: Fakespot was launched for himself, "not doing it just to make money" but to make something many people would benefit from daily [1].
- Treat leaving your environment as the tell for risk appetite; people who avoid risk "will remain in their environment uh probably their whole life and find comfort and solace" in it [1].
- Comfort and legibility are reasons to leave a good job: at Goldman Sachs "everything there was a pattern," from promotion to dress, which is what pushed him to start his own company [1].
- Automate your own manual process: after hours of checking reviewer profiles by hand, he built AI models to "emulate some parts of my brain," one for language, one for purchasing behaviour, one for the company [1].
- Ship the smallest possible surface: Fakespot v1 was a text box, a button, and a trust grade returned in thirty seconds [1].
- Name the project before you start it, and pay for the right name; he bought fakespot.com from a Counter-Strike clan owner for around a hundred dollars [1].
- Startups accelerate character: in the hot seat, building product, team, culture and business at once, "the character develops uh like a rocket ship" [1].
- Obsession is the actual input: post-commute, he worked on Fakespot from 8:30 p.m. to 3 a.m. while employed full time [1].
Media & appearances
- Saoud Khalifah discusses his background growing up in Kuwait and Poland, his early passion for hacking, reverse engineering, and finding security vulnerabilities in software like Microsoft PowerPoint and Adobe products. He explains how he founded Fakespot, a platform using AI to detect fake product reviews, which was later acquired by Mozilla where he became a director. He also reflects on risk-taking, entrepreneurship, and his personal life including family and ongoing projects like deepfake detection frameworks.YouTubeSaoud Khalifah, Building with Obsession, Challenging Comfort ... - YouTube
- SpotifyEp. 139 - Saoud Khalifah: Using AI to Detect Fake Online Reviews
- Interview with Saoud Khalifah [Fakespot] - Venture Bros - Podcast
podtail.com
- Spotify#30 - Saoud Khalifah - #Fil50 - في الخمسينة | Podcast on Spotify
This page shows public professional information only, each fact cited. Is this you? send a correction, or ask for removal within 24 hours, no questions asked.