The stated goal is to reduce drudgery. Do you believe that?
Bloomberg reports OpenAI Looks to Replace the Drudgery of Junior Bankers’ Workload.
My title is a more accurate representation of the true goal.
OpenAI has more than 100 ex-investment bankers helping train its artificial intelligence on how to build financial models as it looks to replace the hours of grunt work performed by junior bankers across the industry.
The group, which includes former employees of JPMorgan Chase & Co., Morgan Stanley, and Goldman Sachs Group Inc., is part of a secretive project inside the startup that’s code named Mercury, according to documents seen by Bloomberg.
Participants are paid $150 per hour to write prompts and build financial models for a range of transaction types, including restructurings and initial public offerings, according to a person familiar with the effort. The company has also granted the contractors early access to the AI it’s creating that aims to replace entry-level tasks at investment banks.
The project underscores the urgency at Sam Altman’s OpenAI to make its powerful AI technology more useful to businesses across a wide swath of industries, from consulting to finance to legal to technology. Despite reaching a $500 billion valuation earlier this month, the world’s largest startup has yet to turn a profit.
The application process for Project Mercury involves almost no human interaction, according to the person familiar with the matter, who asked not to be named discussing non-public information.
The first step is a roughly 20 minute interview with an AI chatbot, which asks questions based on the applicant’s resume. The second phase tests candidates on their knowledge of financial statements. The final stage is a modeling test.
Participants are asked to create their models in Excel and they’re also expected to follow industry norms for formatting the models, including for areas like margin sizes and italicizing percentages.
We all know what the goal is.
What we don’t know is how successful AI will be in that goal.
We also don’t know how this model that hasn’t made a dime supports a $500 billion valuation.
Is AI a Magic Bullet or a a Doomsday Machine?
On October 14, I asked Is AI a Magic Bullet or a a Doomsday Machine?
I don’t believe it’s either but I do suggest these valuations are more than a bit stretched.
For discussion, please see Circular Investment Deals in AI Look Similar to the Dot-Com Bubble
By the way, why can AI just look at some reports and do all of this itself.
AI, like all technology, is a tool which will improve productivity, and create tremendous upheaval in the jobs market. This has been happening for hundreds of years. And each new innovation has accelerated the upheaval in the jobs market.
The overall outcome of the last few hundred years of tech innovation has been a huge improvement in living standards for all.
The downside is how unevenly those benefits are distributed; with those at the top receiving the bulk of the benefits. There are still small improvements for those at the bottom, but the gap is widening.
There is no free enterprise method to better balance the distribution of these gains.
It requires government intervention and appropriate policies. It may indeed require some form of UBI, but I suspect there are many ways to help solve this problem.
For each individual, all you can do is acquire as many skills as possible, work as hard and smart as possible, save, invest and become so wealthy, you no longer need to worry.
So, at best, you will get financial advice that is just as great and useful as what they are providing currently. How great is that? /s
I have yet to encounter a chat bot that can deal with a question. It’s just looking for something similar in a FAQ, but cannot evaluate any real problem.
Oh thank goodness, I thought you said bakers.
“We all know what the goal is.”
The same as moving U.S. factories to China or replacing American IT workers with foreign H1B’s, as Disney famously did. Friend is developing it’s gen 3 humanoid robot. More goals to meet.
ATMs, call centers and online banking cut jobs, yet JPM, Bank of America and other banks opened more brick and mortar branches – friendlier branches, on main streets, in shopping centers, office buildings – with ATMs and “takeout windows” like MCD or DD and safe deposit boxes, to serve their customers.
The past is no guarantee of the future.
the old ladder to money is breaking
…In fact, a third of 2023 grads took up finance roles across several industries, with 64% of MBAs in the sector going into venture capital and private equity. Finance salaries held steady with an overall average of $175k, earning median signing and performance bonuses of $40k and $150k respectively….
All I can say is that I’m glad my career is over. I never had a day out of work never had to claim unemployment benefits had steady, rather than spectacular, progress up the career ladder; retired at 54 when I was quite wealthy. I wouldn’t want to be competing for a job or on a career ladder today or in the future.
Well like the saying goes. Your not building a team your just training your replacement.
Ai is gonna take a lot of jobs. And be just as incompetent as your coworkers.
Was pondering all the tech my employment has added the last ten or so years. Its not really replaced anyone. Just has made how you do your job different. So maybe the ai cuts will like doge cuts to gov then they have to scramble to rehire those who know.
Here’s a guy on Youtube saying the same thing a month ago. Saw another article, sorry, can’t find the link, about 6 months ago saying the same, mid-level financial analysts (Excel jockies,) are doomed.
https://youtu.be/xCrd3ajdlzw?si=BXgHwdXHiQYXh4Po
The question I have about this elimination of “mid-level” and “entry-level” positions is where do high level people come from? From entry and mid level positions. Eliminate those and you kill the talent pipeline. AI is only as smart as what it’s been trained on which is a bunch of other people’s work. When new content creators are no longer created and all the old content creators die off who will teach AI new tricks?
I saw the same thing happen when H1B visa holders were imported to the US and killed off the low and mid level programmers. Companies kept a few old-timers around but eventually became 90% Indian. Some got through the visa / green card / citizenship process (more at the beginning in the 80’s and 90’s,) but most got recycled back to India at some time and a new blank slate came over. Tabula raza. Eventually all the intellectual capital was gone about how everything worked so changes took longer and longer with lower and lower quality.
“AI is only as smart as what it’s been trained on…”
UNTIL it is able to think for itself. This is the point you and many others are missing. Ai is on its way to being able to create new theories, new ideas, new products, etc. on its own w/o human programming or training. This is a lot closer than many imagine.
I would love for you to expand on “it’s a lot closer…” My understanding is that just accumulating a lot of information, is not going to lead to reasoning. Thanks.
A lot closer could be as short as 10 years or as long as 30 years, IMO.
This morning I was playing around with Sora 2 (supposedly the extreme state-of-the art in video creation) and gave it the opening 3-4 lines of a movie script as a test. When you get past the nice visuals, the result was disappointing (gross understatement). Zero character except the painfully obvious.
What AI lacks is imagination. If I had described every detail, it might’ve been better, but then that’s my imagination. My creativity.
What if I’d given Sora nothing, a blank page? Give it almost nothing, no prods. Eg> ‘a room.’
It cycles through its training, and returns a compilation–so far it can’t enhance beyond what it finds. It cannot judge which variation is better? Is really doesn’t know what things mean, or what they should be, if it is not in its training. Maybe it will eventually.
Go back and read my comment about AI infancy.
So you are expecting a qualitative jump?
Don’t you realize that the algorithms and heuristics that AI applies is nothing like those employed by the human brain. It just tries to imitate the result o reasoning, but is often caught out with absurd results.
Do you think that if you just create an infinite amount of beer, one day champagne will flow?
The senior guys will outlast a handful of C suites and their respective golden parachutes, that’s all that matters to them.
@Mish.
Off topic Mish but are you doing an election breakdown tonight? Looks like dems are sweeping elections in many parts.
Oddly enough, Breitbart is covering none of it just whining about Mamdani endlessly.
Democrat Spanberger takes Virginia governorship.
I live in VA. Jay Jones (Democrat) was caught saying awful things about an opponent’s children. He was predicted to lose the VA Attornery General race. However, Decision Desk just called the election in his favor.
Jason Miyares, the Republican running for Attorney General, was to chummy with Trump I guess.
I don’t generally vote so I could care less who wins, but I did think that Jone’s remarks were abhorrent. I guess being chummy with Trump was to much for voters to stomach.
In my view, this doesn’t bode well for Republicans next year.
The headline on Breitbart reads, “Dem Spanberger Beats Anti-Trump Republican Earle-Sears for Virginia Gov ”
As if the republican lost because she was anti-Trump.
The apathy and defeatism sets in over at Breitbart, “So while Democrats will almost certainly celebrate what’s likely to be a smattering of wins for them on Tuesday–and Republicans are left holding out hope for an upset or two or three in different places–the real story is with this deepening trend for Democrats whether or not they can course-correct now that Tuesday’s elections are coming to a close and get their party out of the cosmopolitan gutter that has them geographically limiting themselves before next year’s midterm elections.”
LMAO…Earle-Sears is not anti Trump by any stretch of the imagination.
That’s straight up viper by Breitbart.
According to that left wing rag….Ghazala Hashmi (democrat) wins Lt. Gov. Virginia.
https://www.foxnews.com/live-news/election-day-november-4-2025#&_intcmp=fnhpbt3
Democrat Sherrill large lead in New Jersey Governor race.
New Jersey called for democrat Mikie Sherrill. Large turnout across the board.
NYC Mayor called for Mamdani.
Lol. JD Vance’s brother loses in Cincinnati mayor race.
https://www.pbs.org/newshour/politics/aftab-pureval-wins-reelection-as-cincinnati-mayor-defeating-jd-vances-relative
The winner: Aftab Pureval
People expect Mamdani to round up the Jews, but most of these politicians never deliver.
Does this qualify as a blue wave “mandate” for Trumpers here? Lol.
Who needs the exit strategy now?
There’s no shortage of election coverage on virtually every channel on the TV.
I’m sure Mish will do a recap here, even though it will have no effect on any outcome.
From all the rest of your off-topic posts tonight, it looks like you are having fun rubbing Republican noses in the dog poop. I guess this makes you happy although it will have zero effect on anything.
Zero effect on anything? Lol. It affects everything and if you don’t understand that time to get a check up for dementia.
Of course politics affects everything but my point was that anything Mish or you or I writes has no real effect on anything.
Karl Deninger has a good article about the Mamdani lunacy…
I’d like to know how they are dealing with the hallucinations and errors AI generally produces. Bankers will be dealing with multi-million or multi-billion dollar transactions, if the numbers are off who is liable?
No one has been talking about any liability with any of the errors these AI bots are producing and that wouldn’t be acceptable with airplanes or machinery.
Michal Burry has taken large put positions on a few tech stocks, I may do the same.
https://www.telegraph.co.uk/business/2025/11/04/big-short-trader-makes-1bn-bet-against-ai-boom/
I saw a video somewhere of a robot trashing a kitchen too. So much for robots taking over the world.
The robot only trashed the kitchen because she’d been taken for granted and never listened to.
They are tuning the systems and reducing the occurrences of hallucinations.
How do YOU deal with all the incorrect and misinformation on the internet and in the media (the same thing as machine hallucinations)? YOU learn to discern what is more likely to be true. Not everyone is able to do this. But machines will be able to do so.
Ai machines are still in infancy. Wait until they reach 3rd grade level.
Bakers hallucinate and trash people’s lives even without snorting copious amounts of powder.
Teenage robots almost always leave a mess.
More high tech workers facing the chopping block due to AI also.
All your monies are belongs to us!
Forget money. AI and its worker robots don’t need money. Everything will be provided free of charge to all humans equally.
AI businesses don’t have to make money. They just have to continue long enough to develop AI far enough that it can assume control of human government and then the concept of money flies out the window.
I sense that a lot of people will have problems breaking out of the indoctrination that money equals self-value/worth.
Do you remember DOGE? Do you remember how some members of the public reacted? They trashed Teslas and burned down Tesla dealerships.
And that was just for laying off a few people….
Remember COVID and people being asked to wear a mask? Remember how some members of society wanted to break out into civil war over it?
You keep posting your fantasy dreams of free stuff for everybody, it’s never going to happen. The billionaires will burn down the data centers if it means the slaves will be free.
I don’t think Jojo expects a lot of nice free stuff. More like a survival ration of Soylent Green pellets. Maybe I’m wrong.
I expect Nirvana for all!
Of course, there will be the usual collection of malcontents who cause problems by refusing to accept the rollout of the future. They will have to collected by the robots and deposited into the Soylent Green tanks.
You are a symptom of the old ways that will have to be eliminated.
WTF?? Dude. Relax.
if there’s practically infinite capacity and the billionaires decide to share, then MPO will change his mind. No need to “eliminate”. If those things don’t happen, then MPO’s dour but realistic prediction will come true.
Those who refuse to accept the new reality will cause problems.
In SF books I have read, the malcontents might form revolutionary groups that attack those who who disagree with the majority choice leading to global war between humans and machines
Therefore eliminate was the correct choice of word, What that might entail exactly, I do not know.
I suggest an energy-backed currency. Everyone gets some periodically. You earn some interest on what you don’t spend, so that energy can be invested back into the AI. An exponential wealth tax prevents obligarchy.
I could use a little energy.
How ’bout just a +1? Did that help?
WHY is a currency needed when in a post-scarcity world, everything would be free?
I know this is a difficult concept for the hoi polloi to embrace but think of our current money based, status environment as just one stage that will expire as we move onto the next stage.
Energy is not free to the machine(s). To optimize its use, they will devise their own economic model. They need a currency.
Goods and services provided to humans will cost energy. To ensure some humans don’t demand so much that they crowd out other humans’ demand, a specific amount of the same energy-backed currency should be doled out periodically.
When China and/or AI develops fusion, electricity becomes much cheaper. Still not free. The UBI can turn into a UHI. Still not a “UII” (universal infinite income).
No they do not need a currency. Machines are tireless. They do work 24 hours/day, 365 days/year w/o complaint. They do not get paid and do not need to pay for anything. If more machines are needed for a job, they will be deployed or built.
Also, in the not distant future, we will have unlimited energy via fusion power and the potential of beaming power down to the planet from orbit, tapping into the virtually unlimited power provided by our sun. Energy will be free and abundant.
You clearly do not understand the model I am describing.
AI at the core, is not about true understanding, but gross averaging, meaning it focuses on central tendency +/- one standard deviation. I expect most of the jobs lost will be for IQs between 85 and 115. Jobs requiring IQs between 116-130 and 70-84 will go in the second wave.
Frustration will increase exponentially. Vigilantes will attack AI centers, destroying their energy systems, defecating on NVidia cpus, etc. Police and military will resist the onslaught until they realize they are also being replaced by AI.
This could happen by those who believe humans should not submit to machines and has been depicted in numerous SF books.
But outside of the “humanists” who believe that humans should remain at the top of the intelligence pecking order, the vast majority, finding all their needs met, not having to “work” at anything to earn the ability to pay for a home or food or clothing or anything else, will most likely be happy to kick back and accept all that will be made available to them.
OpenAI will doubtfully screw this up.
I welcome this development, because a UBI or UHI won’t become politically viable until AI hurts people with high incomes. When everyone’s being affected, suddenly UBI will become “common sense”.
I hope AI tells elected officials UBI is a bad idea
Have you asked ?
Maybe ask it what it thinks the challenges are, how it might overcome them, and what it thinks about your POV.
Think I’ll do the same, soon.
I asked my favourite silicon mind. It seemed to like the idea (that I thought of like 20 years ago and hoped even then to see in my lifetime).
Give it a try. Go chat with a silicon mind about it.
Perhaps my adaptation of Niemoller’s poem will help.
—
“Learn to Code”
First the FSD came for the Uber drivers.
But I didn’t drive a taxi.
So I said nothing.
Then the AI wrote the code for the graphic designers,
But I wasn’t “creative”.
So I said nothing.
Then the models came for the junior bankers.
But I didn’t have a Wall Street internship.
So I said nothing.
Then the agents came for the coders and the analysts.
But I didn’t speak Python,
So I said nothing.
And then one day, it optimized my role away —
and I realized there was no one to appeal to;
only a server to petition.
Looks like a good start. Let me know when AI replaces all the Senior Bankers.