Morning Coffee: 24 year-old hedge fund manager’s unhelpful friends. McKinsey & Co.'s weird interviews are coming soon to a bank near you

When Leopold Aschenbrenner’s Situational Awareness fund got into trouble back in June, there were two things which saved him – his strong network of Silicon Valley friends and connections, many of them at Anthropic, and his incredibly valuable pre-IPO stake in Anthropic.  (Well, maybe three things if you count Ken Griffin and Citadel, who took over Situational’s entire public equity portfolio and liquidated it quickly enough to avoid being caught by yesterday's Coreweave meltdown). 

💥Follow us on WhatsApp for news alerts.💥

But 24-year-old Leopold's many friends, with whom he may once have attended Iliad wrestling parties, could be a lot less helpful soon. In fact, they could become unhelpful. 

This is because, for Situational Awareness to realise the value in its Anthropic stake, Anthropic needs to IPO.  When it does, it will be a publicly traded company.  And when you own a big stake in a publicly traded company, the fact that you have so many close relationships with insiders starts to become a compliance headache rather than an edge.

Aschenbrenner can hardly do anything to mitigate his insider status – his Silicon Valley network was the original “situation” which put him in a position where people handed over billions of dollars to a young person barely out of college who had been fired from two previous jobs.  And so, compliance consultants like Greg Worsfold seem to think he will have to restrict his trading in Anthropic (and presumably other companies where he has close relationships).  He might have to observe “blackout periods” around earnings releases and keep a paper trail showing that his investment theses were constructed based wholly on publicly available information.

Of course, this is potentially a problem for the future; Aschenbrenner might not want to sell his Anthropic stake this side of the Singularity.  But it’s one that needs to be taken seriously.  The last time we had a tech stock boom of any similar magnitude, there were plenty of people saying things like “no conflict, no interest” and it didn’t end up well for any of them.

Separately, role playing as an interview technique is rarely used in investment banking.  In other hiring contexts, it’s quite common for people to ask candidates to act out a scenario with their interviewer, to see how they would respond to a client.  The advantage of doing this is that it allows an insight into someone’s personality and ability to think on their feet. 

The disadvantage is that it is often really weird and embarrassing for the interviewer, so many bankers will flat out refuse to do it.  (The only exception we’re aware of is David Bokman, the co-head of family office resources at Morgan Stanley, whose team uses roleplaying in order to see whether wannabe private bankers have the emotional intelligence to deal with the extremely rich).

But what if you could get an AI chatbot to take on the task?  This is being trialled by McKinsey & Co.  From this autumn, entry-level candidates to the consulting firm will have a ten minute conversation with an “avatar” that has been trained to imitate stereotypical client personalities.  Disappointingly, it doesn’t seem to have been given a cutsey name yet, but apparently it is not meant to provide “gotchas”.  It is just meant to test candidates’ abilities to notice important information and to think rather than giving prepared answers.

It’s highly likely that this will soon be coming to banking interviews too, and when it does it might be in a significantly more aggressive form.  Sales and trading desks might be interested to find out how a candidate handles someone who is angry about a trading error, while capital markets teams might want to grade potential employees on their ability to tell a CEO that their IPO is pricing lower than expected.

What is less likely is any use of this sort of technology for anything other than entry-level roles and internships.  As a way of quickly sorting through a large potential talent pool, the AI roleplay might be a very good way to structure the kind of interactions that make it easy to tell a potential future star from a dud.  But there is a steep price in dignity and awkwardness. In the future, one way to measure the progress of your career will be by whether or not anyone would dare to ask you to do a role-playing exercise with an AI.

Meanwhile …

The $1.2bn fine which Bank of America got hit with seems to be quite disproportionate to any actual offences.  The bank was fined for technical breaches relating to poorly designed alerting systems, but the penalty is larger than comparable cases in which large amounts of criminal money was provably laundered.  It almost looks like a political vendetta by President Trump against Brian Moynihan. (Bloomberg)

One of the things which senior investment bankers find almost as awkward as being asked to role-play at interviews is being asked to drum up clients for the wealth management side. Although “One Bank” sounds like a great slogan, it is just very difficult to switch between portraying oneself as a trusted advisor and trying to do a bit of cross-selling.  Citi’s rainmakers are the latest to get the “just one client each” speech. (Financial News)

Even with the benefit of the Situational Awareness trades, Citadel’s equities funds have been outperformed by its “tactical trading” (quant) unit. (Rupak Ghose)

The Perella Weinberg / Piper Sandler merger talks have ended.  Apparently “a critical mass of Perella Weinberg’s top talent did not get on board with the deal” and the two boards couldn’t agree on “key financial and personnel matters”.  (FT)

Antoine Iskander had been in talks with Millennium about spinning his pod out into an independent manager.  Those might have been complicated by some losses this year, and now he’s going to go to Balyasny instead. (Bloomberg)

It’s often a fun practical joke to drop a made-up name into conversation and see if the trading floor’s biggest blowhard claims to have worked with them or met them at an event. Apparently this works in tech too, where 53% of men surveyed claimed to have heard of AI pioneers “Craig Dalton” and “Susan Milner”, despite the fact they don’t exist. (WIRED)

Follow me on X. Follow me on LinkedIn. 

Have a confidential story, tip, or comment you’d like to share? Contact: +44 7537 182250 (SMS, Whatsapp or voicemail). Telegram: @SarahButcher. Signal: sarahbutcher.22  Click here to fill in our anonymous form, or email editortips@efinancialcareers.com. 

Bear with us if you leave a comment at the bottom of this article: comments are moderated intermittently by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. You must take sole responsibility for comments you post on this site. We will take reasonable steps to weed out anything that we consider to be offensive or inappropriate.