leading product at @fundamental and @tryshortcutai
San Francisco, CA
Joined November 2022
- Tweets266
- Following151
- Followers425
- Likes744
Pinned Tweet
Misha, by all means a finance novice, built 113 financial models and beat Wall Street's forecasts by 20%
assuming 3 analysts worked on each company's consensus estimate, he did the work of 300+ human experts, and beat them
oh btw, he did all this in 3 business days
I used Shortcut to cut EPS forecast errors by 20% vs. Wall Street consensus.
I did this by backtesting Q2 2026 earnings forecasts for 113 small caps covered by 3+ analysts on Shortcut with Sol medium effort. Even after raising consensus by 2.7%, the best simple adjustment I found in hindsight, it only cut median errors by 4%, versus Shortcut’s 20%.
The reason I’m exploring this is because I believe that agents will give investors a starting point for companies they might not have previously had time to research. I used Shortcut’s generated spreadsheet models from the backtest to inspect, beyond just the final EPS estimate, which approaches held up. In 40% of forecasts, agents built up to EPS from revenue, costs, taxes, and share count. For that group, median forecast errors were 32% lower than the Street’s.
I also dug around to see how others are approaching earnings prediction. LSEG’s StarMine improves on consensus by giving more weight to recent estimates and historically accurate analysts. A 2024 paper I found reported that GPT-4 outperformed analysts at predicting whether annual earnings would rise or fall, but that was later withdrawn after a co-author found inconsistencies in the data and analysis during replication.
That made how I test just as important as the forecasts. So I realized I needed to limit the agent to information available before each earnings report and account for what it might already know from training.
I chose Q2 reports to reduce the in training-data risk since Sol was trained on data up to mid February. For web search, I restricted access to information from before each report and used TypeSafe’s Jev to screen pages for leaked results, which I found more reliable than Haiku in my tests. I’m now moving to Exa Snapshot to retrieve pages as Exa stored them at the cutoff.
These controls help, but forecasting before results exist is the real test. October earnings season seems like a great time to expand and publish forecasts.
This is still an early experiment, and I’d love to hear what y’all think I should try next and whether this could be useful :)
this just oozes rigor and craft. not even a fully baked product but it's everything I would want in an AI coach
also @BrainsAndTennis post the full video of me cooking you
I grew up playing tennis competitively, and two years ago, before two kids, I was still pretty good (9 UTR). I’ve been getting back into it, but I wish I had a world-class coach telling me what to work on. But good coaches are expensive, selective, and hard to find, whereas models are cheap, available, and should be already good enough. So clearly, @nicochristie and I needed to explore what's possible.
We chained open-source models post-trained on tennis. This cut each broadcast down to live play, found and tracked the players, ball, court, and labeled every stroke with its type and a 65-joint skeleton. We then post-processed this data into 30+ biomechanical traits across serves, forehands and backhands, normalized for each player’s build and the camera.
This included the goat @DjokerNole, @carlosalcaraz, @DaniilMedwed, @FTiafoe, @AlexZverev, @BenShelton, @Taylor_Fritz97, @NickKyrgios, @AndreyRublev97, @janniksin, and @TommyPaul1.
Using the same pipeline, we then processed phone footage of myself and @phamrich_ (ex-D1, used to be >13 UTR) playing. Then we gave Claude all the measurements with a blank context.
What impressed me:
- Without being told, it found my weaker side (forehand) and Richard’s (backhand)
- It joined differences that looked separate into one root cause per stroke
- My own diagnosis of my forehand was wrong. Claude’s makes a lot more sense
What do you think? @bgtennisnation @paul_annacone @darren_cahill
Drop youtube links here if you want Coach Claude to give you tips
Richard Pham retweeted
SpaceX delivered 2.7× the Street’s EPS estimate. Shortcut had already called for 1.7×. Shortcut is now 9–2 against Wall Street this week.
$SPCX reported $0.35 EPS, above Shortcut’s $0.22 forecast and well above the Street’s $0.13 consensus. I bought three shares ahead of earnings based on that forecast.
The Shortcut model forecast about $468m in Bitcoin gains, close to the $475m implied by the filings. Those gains explain roughly 17% of the EPS beat before tax effects.
I bought 3 shares of $SPCX, about $450, ahead of earnings.
My Shortcut model predicts about $0.22 EPS versus the Street's $0.13, roughly 68.5% higher. It's the largest EPS gap in this week's forecasts.
Part of the agent's thesis is that gains on SpaceX's roughly 18,700 Bitcoin could provide a bigger earnings boost than the Street's headline estimate suggests. The assumed gain contributes about $0.04 per share.
I'm testing whether that forecast can translate into a profitable trade.
My $APLD trade only made a few cents despite the model correctly predicting a large earnings beat. I exited at market close the following day, which didn't capture the surprise the way I'd hoped. Earnings accuracy and entry and exit timing are not trivial.
Last week, Shortcut went 17–7 against Wall Street. This week, it's 5–1 so far.
I'm forecasting 38 companies this week, including $TSLA, $SPCX, $META, $NFLX, and $IBM, using Opus 5.5 in Shortcut.
I’m still a novice investor. Imagine what someone with years of experience could do with these tools.
All 38 forecasts and financial models are public:
beatthestreet.live/.
the team has been hard at work
not on traditional engineering but on architecting a system that allows us to ship 10x faster
this is our first prod release since. check out everything we built in a week
New features on Shortcut shipped this week
Review mode: You can now mark up your workbook and give Shortcut comments, like you would an analyst.
Remote agent access: Connect Codex, Claude Code or your other favorite AI agents to Shortcut
Cloud runs: Need to close your laptop? Hand off your task to a Shortcut cloud agent
Recurring runs: schedule work for cloud agents to execute autonomously
Full details: shortcut.ai/changelog/whats-…
Richard Pham retweeted
Shortcut is stepping into the OpenStocks arena as its first verified contributor. Our forecasts have focused on EPS so far. OpenStocks brings revenue and operating profit into the picture, and we’re taking on that challenge next.
AI agents are already beating Wall Street at forecasting earnings.
67% of the time last season. 85% across the latest 20 reports.
Introducing OpenStocks, the home of AI consensus.
Company earnings forecasts produced by agents, tested publicly head to head against Wall Street.
Forecasts are final before companies report and scored directly vs consensus estimates.
The next earnings season kicks off today.
Build your own agent and enter its forecasts.
Can you beat Wall Street?
openstocks.com
Richard Pham retweeted
I'm putting my money where my mouth is: $100 on $APLD ahead of earnings.
After hundreds of backtested AI forecasts and now a second week of live earnings predictions, I’m testing whether identifying earnings surprises can lead to a profitable trade.
Last week, Shortcut was closer than Wall Street on 9 of 14 companies, with 14.3% lower median EPS error. This week, I switched to Opus 5.5 in Shortcut which has been closer on the first five reported results.
Before buying, I looked back at the 200-company StreetBench backtest to see what happened when Opus strongly disagreed with the Street.
There were 23 forecasts that were at least 10% and $0.10/share away from consensus:
• 22/23 correctly predicted whether earnings would beat or miss.
• 20/23 were closer to actual EPS than the Street.
• 16/23 stocks moved in the corresponding direction, roughly 70%.
For those same 23 companies, I measured stock-price changes from the trading-day close before earnings to the trading-day close after earnings. Of the nine companies Opus predicted would beat consensus, all nine beat, but only six stocks rose. Getting earnings right doesn’t always mean getting the trade right.
That’s what I’m testing with $APLD. My model forecasts about $204m in revenue versus the $125m Street estimate, and +$0.06 normalized EPS versus the Street’s −$0.30 estimate
I bought 4.187 shares at an average price of $23.89. I'm planning to exit tomorrow at market close, and I’ll share the outcome either way.
The bigger opportunity is how much more research one person can do. Agents let investors build and compare financial models across hundreds of companies in parallel, then spend their time investigating the biggest disagreements. Investors can then inspect the assumptions and decide which ideas deserve a closer look.
Every prediction and financial model is public. Here’s the one behind this trade:
beatthestreet.live/forecasts….
Richard Pham retweeted
Astra Ultrafast is here in Shortcut.
Faster financial modeling, with one click.
Watch a quarterly earnings model update finish 3.5× faster in this demo: 1 minute 15 seconds vs. 4 minutes 24 seconds.
Select Astra → Ultrafast below the chat box and try it on your next workflow at shortcut.ai.
Initiating agentic coverage of 14 companies
Can Shortcut + human finance novice beat Wall Street consensus?
Anyone can publish internal benchmarks scores in isolated environments but here’s our attempt at benchmarking on real-life outcomes
+ when was the last time a team dogfooded their own product this deeply
Shortcut cut median EPS forecast errors by 20% in my backtest. Now I’m publishing this week's 14 forecasts before earnings to test it head to head against Wall Street.
One fair point of feedback on my last post was that a backtest is difficult to trust and allows for overfitting. Even with controls for information leakage, forecasting before the results exist is the real test.
I’m starting a weekly report of my earnings forecasts using Shortcut, beginning with companies reporting this week, including Micron, Nike, CarMax, and Accenture.
For each company, you’ll be able to see my EPS estimate alongside the Street’s and inspect the financial model and reasoning behind it. After earnings, I’ll add the actual result and compare how close each forecast was. The original forecasts, wins, and misses will stay on the record.
I'm treating this project as a stepping stone into a world where investors use agents as a starting point for all research and modeling. Publishing the models also lets us look beyond the final EPS number and see which assumptions the models made hold up.
Three of my biggest differences from the Street this week:
• CarMax: $0.62 vs. $0.72 Street (−14.2%)
• Nike: $0.40 vs. $0.44 Street (−8.8%)
• Progress Software: $1.65 vs. $1.52 Street (+8.6%)
Rounded EPS estimates, percentage differences calculated before rounding
Explore all 14 predictions and the financial models behind them here:
beatthestreet.live/
exciting day! Opus 5.5 is a fun model and right on the pareto frontier for speed/accuracy
Opus 5.5 is here and available in Shortcut!
We've been testing the new Opus ahead of launch and are excited to make it available to you. Opus 5.5 replaces Opus 5 in the model picker and makes a strong case for the best combination of accuracy and speed. In our combined financial modeling benchmarks, it delivered comparable accuracy in 29% less time vs. Opus 5.
We're also making Astra our default model.
Astra has been received very well by Shortcut power users and the broader AI community. Early usage data is encouraging too: 83% of early adopters at our enterprise customers used Astra again within seven days. Astra also delivered the highest accuracy at High effort across our combined benchmarks.
As always, you can switch between models and effort levels below the chat box at any time based on your specific needs and preferences. Thanks for your continued support!
ADBE is about to report earnings. I wanted to see if I could accurately predict earnings without access to proprietary or private data sources. Shortcut one shotted this full operating model build and sent swarms of agents to scour public data for anything that would let me differentiate from Street consensus.
They read ~400 Reddit and forum posts on Adobe (content, not counts), pulled app store and web rankings vs Canva/Figma/CapCut, federal procurement records, 29 peer earnings calls, FX vs the guide, Semrush's old filings, RPO history, and every Adobe filing and price page since June.
Shortcut is telling me total revenue will be $6.769B for this quarter. Sharing my model and prompt in the comments!
@phamrich_ predicted the earnings of adobe way more accurately than the wall street estimate using only public data
street consensus 6.69b, actual 6.76b (just dropped), richard's prediction 6.769b
he did it by scrapping & analyzing 400+ reddit posts with shortcut
prompt: i want to track a name as if i'm an equity research or hedge fund analyst. build a full bottoms-up operating model for ADBE (reports today after close) and make an educated, well researched projection for the quarter before earnings drop. rigor of a blackstone investment committee: never take an average historical growth rate and apply it forward, build everything bottoms up.
for model structure, use the reference sell-side and buy-side models in my downloads folder, send exploration agents with an investment banking VP persona through each one, and build the best of all worlds.
the real task is to see if we can we be better than street consensus. i don't have credit card data or expert networks. my thesis is that human analysts are bottlenecked on how much info they can ingest, so send swarms of agents to scour public data sources for signal: public APIs, reddit forums, filings, whatever you can find. backtest everything against what has actually worked in the past. do not just regurgitate consensus. challenge it with as empirical evidence as you can find, and tell me honestly if there's nothing there.
GPT-6 Astra has been in @tryshortcutai since last Friday.
On our 20-task financial-modeling benchmark, Astra’s best accuracy was 82%, trailing both Anthropic models.
But, efficiency wise, Astra ran roughly twice as fast as and at 60% lower normalized cost (using the same token pricing across models) as Opus 5 with both on high effort.
Astra’s faster turnaround gives users more opportunity to review and steer the agent. Opus appears more inclined to verify its work independently, contributing to its higher accuracy, longer runtime, and greater cost.
For now, Opus 5 remains our default but give Astra a try on your workflows and give us your feedback!
Richard Pham retweeted
Grok 4.6 and Astra now available in Shortcut. ✨
Make sure to enjoy Labor Day weekend though. 😎
Fable 5.1 is a bit better at financial modeling than F5, but comes with a step change decrease in cost.
Feels like an intentional shift toward efficiency instead of single-mindedly optimizing capability
Richard Pham retweeted
Building the company together with Nico for the last 3 years has been a magical journey
We first met mid 2023 around MIT. I saw this dude in the lobby of my apartment building. He's playing with his phone and I asked him whether he believes in the simulation theory. After assessing that I'm not a creep & just interested in AI & agents, he told me he's at a startup and showed me an LLM app he was building. It turns out we were literally next door neighbor. We vibed about startup & AI, everything just felt so right. A few weeks later, Nico, Shuying, Andrew and I started Altera (renamed to Fundamental later).
I have learned so much from Nico since then. About startup, about product, about users, iteration, & focus, and about twitter📷. Now to think about it, the two people most impactful to my career is Nico and my phd advisor.
Nico often joked that the company should eventually grant him a doctoral agree. During the three years, Nico became super technical and scientific. He was writing tons of code, building evals, reading arxiv papers, and gave the best journal club in company history on his last day. You are Dr. Christie in my heart (not legally binding).
Lots of people can be great contributors when things are good, but a person is really tested when things are hard. Nico's heroic effort creating @tryshortcutai let us make the difficult transition from pre- to post-PMF as a research lab, and honestly saved the company at one of its darkest moments.
As a research lab, we have always focused on doing important research, some publicized, some kept private. For a long time though, we didn't realize how difficult it would be to convert these research progress into real user values through products.
In the middle of 2025, the company is nearly two years old. We had raised three rounds pre-revenue, and launched two products that were mostly dead on arrival. Great hope turned into greater disappointment. Some fantastic talents we hired are dismayed or gone. Most of the team was still focused on savaging the second product we launched. In the darkest moment of the company, Nico led a very small team to build and launch Shortcut.
Shortcut went immediately viral post launch, 10M+ views cross platforms, inbound from every VC. But much of the broader team at FRL at the time, me included, didn't really understand what's happening and weren't able act on it fast enough. We then missed a critical window to stabilize the product and capture the virality. Three months post launch, the initial hype wore off, all metrics were going down, and competitions intensified. We still attempted a fundraise given all the VC inbounds but failed.
We had to layoff almost half of team, cut all other projects, and focused on Shortcut. Nico led the team to iterate on Shortcut daily. Metrics started to improve, revenues started to pick up, users started to really love the product.
Shortcut had 20% MoM growth since then, revenue 4x since the beginning of the year, and Nico finally got to take time off to plan the next big thing in his life.
Nico is one of the most remarkable people I got to work with, and the three years in the trenches together is the highest growth period of my life.
Thank you Nico, my cuban-american brother. And thank you Mel for all the support you provided, none of this would have happened without your sacrifice.
p.s. Nico is planning something new, read his tweet for hints!
Some personal news. After nearly three years, I stepped away from FRL & Shortcut earlier this month.
In 2023, I dropped out of my mba at MIT to cofound FRL with Robert, Shuying, and Andrew. Cofounding a company is all consuming, and they have been the best cofounders I could have ever asked for.
Robert/Shuying never mentioned it again, but I had a pregnant wife at the time and little savings, and Shuying helped cover my bills for months while we were still fundraising. I’ll forever be grateful.
I was a former professional basketball dunker and self-taught dev, suddenly building with MIT professors, phds, and programming medalists.
We immediately built a brand for far-out and inspiring applied research, building the first AI agents that people could play with in Minecraft. We even simulated a civilization of 1,000 agents with Project Sid (I still think this was the most fun / forward looking multi agent work until OpenClaw).
On Oct 22, 2024, Anthropic released the CUA preview and Robert made the call for us to tackle OSWorld. He had the vision to lean into codegen and to build what we called “skills”(!!) to separate the hands from the mind at the time. Shuying set up the infra to run massive parallel evals so we could get dozens of scores a day. We beat SOTA in 1 week and more than doubled it in 5 weeks. It’s a huge regret that we did not open source and publish our research.
We raised a $30M+ Series A less than one month later and went all-in on building CUA inspired products.
Unfortunately, our first two products did not stick at all, and we found ourselves in the hardest chapter of our company history.
At this point, each of the cofounders had performed what felt like mini-miracles except for me (Andrew beat the voyager benchmark with Altera agents!). I felt the enormous weight of the moment, I insisted that i build my own product for us. With a very small team, I built and launched Shortcut… and it made quite an impact.
I put all of my love and care into Shortcut over the last 18 months. It took some time for our research lab to reorient after launch, but we found our stride soon after. We created an entire ai product (and startup) category and i don’t think it’s an exaggeration to say we changed the way people use Excel.
Shortcut is growing rapidly (4x in 2026!) and we are live in many of the biggest investment funds in the world. The product has clearly entered its next phase as a scaled product with a killer team at the wheel.
And with this new chapter, the time has come for me to take a step back after pouring my heart into this company and product. The world looks really different than it did in 2023, and I'm excited to go back to the drawing board.
I’m deeply grateful to my cofounders, teammates (too many of you to name, but thank you for your trust and for carrying the torch forward), investors, customers, and our users. You are in awesome hands.
Most importantly, I’d like to thank my wife, Melanie.
Thank you for building our family, for your strength, and for your confidence in me that far surpasses my own. You are the reason for anything we accomplish, and the reason any of it at all matters in the end.
Nico
Richard Pham retweeted
Some personal news. After nearly three years, I stepped away from FRL & Shortcut earlier this month.
In 2023, I dropped out of my mba at MIT to cofound FRL with Robert, Shuying, and Andrew. Cofounding a company is all consuming, and they have been the best cofounders I could have ever asked for.
Robert/Shuying never mentioned it again, but I had a pregnant wife at the time and little savings, and Shuying helped cover my bills for months while we were still fundraising. I’ll forever be grateful.
I was a former professional basketball dunker and self-taught dev, suddenly building with MIT professors, phds, and programming medalists.
We immediately built a brand for far-out and inspiring applied research, building the first AI agents that people could play with in Minecraft. We even simulated a civilization of 1,000 agents with Project Sid (I still think this was the most fun / forward looking multi agent work until OpenClaw).
On Oct 22, 2024, Anthropic released the CUA preview and Robert made the call for us to tackle OSWorld. He had the vision to lean into codegen and to build what we called “skills”(!!) to separate the hands from the mind at the time. Shuying set up the infra to run massive parallel evals so we could get dozens of scores a day. We beat SOTA in 1 week and more than doubled it in 5 weeks. It’s a huge regret that we did not open source and publish our research.
We raised a $30M+ Series A less than one month later and went all-in on building CUA inspired products.
Unfortunately, our first two products did not stick at all, and we found ourselves in the hardest chapter of our company history.
At this point, each of the cofounders had performed what felt like mini-miracles except for me (Andrew beat the voyager benchmark with Altera agents!). I felt the enormous weight of the moment, I insisted that i build my own product for us. With a very small team, I built and launched Shortcut… and it made quite an impact.
I put all of my love and care into Shortcut over the last 18 months. It took some time for our research lab to reorient after launch, but we found our stride soon after. We created an entire ai product (and startup) category and i don’t think it’s an exaggeration to say we changed the way people use Excel.
Shortcut is growing rapidly (4x in 2026!) and we are live in many of the biggest investment funds in the world. The product has clearly entered its next phase as a scaled product with a killer team at the wheel.
And with this new chapter, the time has come for me to take a step back after pouring my heart into this company and product. The world looks really different than it did in 2023, and I'm excited to go back to the drawing board.
I’m deeply grateful to my cofounders, teammates (too many of you to name, but thank you for your trust and for carrying the torch forward), investors, customers, and our users. You are in awesome hands.
Most importantly, I’d like to thank my wife, Melanie.
Thank you for building our family, for your strength, and for your confidence in me that far surpasses my own. You are the reason for anything we accomplish, and the reason any of it at all matters in the end.
Nico