> because this is marketed as being for macro investing, I would expect to see a level of rigor and quantitative analysis consistent with that.
Thanks for bringing this up - while we talk about Soros' forecasts and comparing them against those of an LLM, in the end Soros is not a forecasting tool, it's an analytical framework.
There is a gap between quant modeling and geopolitical analysis that we seek to fill. Specifically, quant models are great at capturing statistical regularities in financial time series but typically treat geopolitical shocks as exogenous noise. Meanwhile, geopolitical analyses in the policy and intelligence communities (with the exception of Bueno de Mesquita [BdM]'s work) provide deep contextual reasoning but rarely produce probabilistic scenario structures or asset-level transmission mappings that can directly inform capital allocation.
We will be shortly publishing a technical preprint laying out the Soros framework in full, but the TL;DR is: we model geopolitical events (or crises in the literature) as partially observed ("fog of war") stochastic games with multiple actors jostling for control over resources. We map out actors across various axes (think of these as actor embeddings), identify key decision points, and enumerate paths across them to estimate scenario probabilities. The scenarios in turn have associated transmission flows and market implications. We will evaluate those as mentioned in the sibling comment. Happy to discuss more.
First, thank you so much for signing up to try out Soros!
You are absolutely right, of course, to ask about accuracy. TL;DR: we don't have any formal calibration data yet.
The reason why is interesting, though, and it strikes at the heart of global macro investing in particular: things change, often, and sometimes dramatically. Basically, geopolitical "events" are really smeared across time (and sometimes space). Each event update can lead to a cascade of new scenarios branching off and older ones dying out, each with implications on capital flow. It's difficult to disentangle, which is why our preference has been to enable the system itself to monitor feeds, but also update its alerts as it deems fit, and re-run the analysis when it feels there's been enough of a change of state (pun not intended).
One markets-focused eval we have been building towards (and apparently you have been thinking of as well) is comparing against LLMs. Our plan is to run simultaneous comparisons against a variety of frontier models, armed with the same information that we provide Soros, but without the structural framework and simulation engine we've built though. Ideally we want to map out the Pareto frontier of model capability vs realized returns, and examine performance over horizons, asset classes, and so on, and have concrete numbers on where Soros pushes the curve outwards.
This is being built :), and we hope to get there in the coming few weeks!
* This brings us to a larger question - why did we build Soros?
First, let's address the elephant in the room: we were inspired by George Soros' theory of reflexivity and how human tendencies affect markets more prominently than expected. Yes, there's a corny backronym [0]. No, this is not a political statement or endorsement of his views.
Coming back to the main point, we (the founding team at Lookback Labs) have both spent a long time at the intersection of financial markets, technology, and machine learning. During that time, one key thing that kept bothering us [1] was simply this: when a geopolitical crisis breaks, an investor's actual problem is not really to find out "what is happening now" — it's more of "which scenario plays out, how likely is each one, and what do I buy, sell, or hedge under each? For how long?"
There are a ton of existing tools and services that seek to answer the first question reasonably well (newsletters such as StratFor, publications such as Foreign Affairs and Foreign Policy, Bloomberg terminals for breaking news, etc.).
None of these answer the other questions particularly deftly. Sure, one can engage with ChatGPT (or Claude if one prefers), and play through multiple scenarios. You will, of course, miss out on the grounded structural model that powers Soros' analysis, along with the simulations that serve up the relative probability estimates.
Also, one of the worst things purely LLM-based ad hoc frameworks do is assume that countries are monolithic decision-making units from a game-theoretic perspective. This is hardly the case - "Iran" doesn't make choices, Mojtaba and the IRGC faction does. "China" doesn't decide, the Politburo Committee does. And so on.
There are of course formal analytical frameworks that dig deeper, studying groups, factions, organizations that are jostling to gain control (Bruce Bueno de Mesquita's Expected Utility Model and selectorate theory [2] is the most academically serious and is a prime inspiration for our system design), but they are extraordinarily hard to operationalize in real time, and produce no market implications.
To sum up, the choices are stark: ask AI and hope for the best, or build out your own systematic framework to organize evidence, assumptions, and implications. We chose the latter path.
Zooming out, our mission at Lookback Labs (https://www.lookbacklabs.com/) is to build "the intelligence layer for AI-native investing"; accordingly, Soros is the first of several agentic systems that we are designing across the systematic and discretionary spaces, that are both usable and useful from the get go, and not merely demo eye candy.
* Some minor details:
(1) We are currently in private beta for Soros and are onboarding selectively.
(2) The static demo is not completely static; you can still chat with the analysis (up to 20 messages a day per IP).
(3) We are still working on pricing: something that captures the value Soros provides.
(4) We want this to work for individual investors as well, not just institutional desks, and would love to price accordingly.
We're curious to hear what the HN community thinks about our approach. AUA!
Feel free to reach out offline if you'd like! We are, sadly enough, on LinkedIn, but are also available via email (anshuman/karen@lookbacklabs.com)
PS: As is probably obvious to the diligent reader :), every token in this post has been lovingly handcrafted by the Lookback Labs team.
[0] Scenario-Oriented Reasoner for Opportunity Synthesis. Lol.
[1] Many things bothered us. Buy us drinks, get stories.
[2] We heartily recommend two of BdM's books: "Predicting Politics" and "The Dictator's Handbook"
Hmmm, good question. I think one interesting incident for us was when we saw scenario probabilities being updated near last Friday EOD for the US-Iran conflict, biased towards further kinetic action by the US around Kharg island (?). This was basically captured from changes in odds for Polymarket events that the system was tracking.
The news came in a few minutes later, post equity market closing.
> So someone monitoring Polymarket could have reached the same conclusion?
Maybe? If they are professionally trading prediction markets, I'm pretty sure that would be the case. Polymarket especially is a great source of insider traded information, as you pointed out.
We do near realtime tracking of most major markets, plus X accounts that Soros identifies as being important. The system also composes search queries per analysis, along with frequency of scanning, and that's run as requested. (We use a mix of Perplexity and other smaller search providers, along with Exa via OpenRouter's integration.)
Other inputs might be direct statements from leaders involved in the conflict, especially Trump. Also maybe bond market and oil market price movements?
Then you would want to generate an alert when you an actionable prediction. You don't want the user to have to prompt the AI. It needs to be running in the background, having been prompted on the scenarios to monitor?
Responding just because it's a pet peeve of mine: Cixin Liu did not invent the dark forest hypothesis. People were discussing it, and writing science fiction books about it, for decades before the 3BP books were published. Nothing against him, and he definitely helped popularize the concept, but I think it's incorrect to refer to it as "Cixin Liu's hypothesis".
Was curious as a lover of the 3BP series, google gave me this:
"We've been sitting in our tree chirping like foolish birds for over a century now, wondering why no other birds answered. The galactic skies are full of hawks, that's why." (The Forge of God, Legend edition, 1989, pg 315).
The Forge of God and its sequel, Anvil of the Stars, are amazing books for anyone interested in the dark forest theory, by the way. A bit slow and contemplative, so you have to be in the right mood, but they're one of my favorite reads of the last few years.
I think there's a passage that even uses an analogy of a forest, though I'm not sure.
Just like Amerigo Vespucci put the name "America" on a map and people starting referring to the New World as such, although he didn't discover it himself.
> I feel the more of these services come to being, the more likely it is that every website starts putting up gates to keep the bots away
That's why we need microtransactions, because I'd rather be able to have both nice AI services and useful data repositories that they pull from, than have to choose just one. (and that one would be AI services, because you can't stop all the scrapers, so data sources will just keep tightening their restrictions)
All financial transactions have a fraud risk. Microtransactions are no different. But any microtransaction system faces a choice: continually pop up payment confirmations (unusably annoying), or automatically accept charges (vulnerable to fraud).
Click fraud on adverts is a form of microfraud, and pay-per-click is the existing form of microtransaction.
There's zero evidence that this exists, if only because there's very few examples of working microtransactions systems at all. Adverts are not micro transactions. (I don't care how you want to use that word - nobody else uses it like that)
But, all of the systems that I've seen (Blendle, video games) have had no problem at all with fraud, and a very small amount of annoyance to value delivered.
There's simply no reason to believe that this will be a problem, either empirically or theoretically.
"I'm going to build a system for transferring money, and I am confident that nobody will ever try to defraud it" -- someone who is about to lose all their money
- pay before viewing: how do you know that the thing you're paying for is the thing that you're expecting? What if it's a rickroll or goatse?
- so do you give refunds a la steam?
- pay and adverts: double-dipping is very annoying
- pay and adverts: how do you know who you're paying? A page appears with a micropayment request, but how do you know you've not just paid the advertiser to view their ad?
- pay and frame: can you have multiple payees per displayed page? (this has good and bad ideas)
- pay and popups: it's going to be like those notification or app install modals, yet another annoyance for people to bounce off
- pay limits: contactless has a £30 limit here. Would you have the same payment system suitable for $.01 payments and $1000 payments? How easy is it to trick people into paying over the odds (see refunds)?
- pay and censors: who's excluded from the payment system? Why?
> If it was that easy, it would have been done.
Part 2: business model problems!
- getting money into the system is plagued by usual fraud problems of card TX for pure digital goods
- nobody wants to build a federated system; everyone wants to build a Play/Apple/Steam store where they take 30%
- winner-take-all effects are strong
- Play store et al already exist, why not use that?
- Free substitute goods are just a click away
- Consumers will pirate anything no matter how cheap the original is
- No real consumer demand for micropayments
=> lemma from previous 3 items: market for online goods is efficient enough to drive all marginal prices to zero
- existing problem of the play store letting your kid spend all the money
- friction: it would be great if you didn't have to repeatedly approve things, such as a micropayment for every page of a webcomic archive. But blanket approval lets bad actors drain the jar or inattentive users waste it and then feel conned
- first most obvious model for making this work is porn, which is inevitably blacklisted by the payment processors, has a worse environment for fraud/chargebacks, and is toxic to VCs (see Patreon and even Craigslist)
- Internet has actually killed previously working micro-ish payment systems such as Minitel, paid ringtones (anyone remember the dark era of Crazy Frog?); surviving ones like premium SMS and phone have a scammy, seedy feel.
- accounting requirements: do you have to pay VAT on that micropayment? do you have to declare it? Is it a federal offence to sell something to an Iranian or North Korean for one cent?
> "I'm going to build a system for transferring money, and I am confident that nobody will ever try to defraud it" -- someone who is about to lose all their money
You seem to have conjured the impression that micropayment systems have to be radically different than current payment models, which is wildly mistaken.
You can build an effective micropayment system using only currently available tools (digital wallets, microcurrencies, digital storefronts, review systems) that have most/all of the nice properties of existing platforms, which invalidates almost every single point you make.
Few of these points seem very well thought-out - they're mostly relatively easily refuted by using logic and/or pointing to what the industry is already doing.
> pay before viewing: how do you know that the thing you're paying for is the thing that you're expecting?
In the exact same way as current digital storefronts.
> How easy is it to trick people into paying over the odds
What are "the odds"? Are we betting now?
> so do you give refunds a la steam?
Yes, exactly like current digital storefronts.
> - pay and adverts: double-dipping is very annoying
Exactly like current digital platforms (e.g. Spotify, YouTube premium).
> - pay and adverts: how do you know who you're paying?
What does this even mean - how do "adverts" factor in to "how do you know who you're paying"??
> pay and frame: can you have multiple payees per displayed page?
What does this mean??
> - pay and popups: it's going to be like those notification or app install modals, yet another annoyance for people to bounce off
A theory that is trivially dispelled by empirical evidence of the tens of billions of dollars in microtransactions that US players spend on free-to-play games. You create a microtransaction wallet currency that is roughly equivalent to normal money, and then you pay for things by clicking on them, like in a normal game with microtransactions. Empirically, people get used to this very quickly and the friction becomes unnoticeable.
> - pay limits: contactless has a £30 limit here
What does any of this have to do with contactless payments???
> Would you have the same payment system suitable for $.01 payments and $1000 payments?
It's pretty easy with a few seconds of thought to think of systems that handle both of those cases well. For instance, you can make it so you have to hold down a button to purchase something with your microcurrency, with the duration of the hold (nonlinearly) proportional to the cost of the item.
> - pay and censors: who's excluded from the payment system? Why?
Exactly the same as current platforms - the platforms/wallet providers determine that.
>> If it was that easy, it would have been done.
Objectively false. There are many good ideas that have failed because of market factors or poor marketing. In this case, the prevalence of ads, and people's generosity at a time before scraping has truly taken off, is subsidizing the market. The generosity will decrease and doesn't scale, and I shouldn't have to point out the problems with ads.
> - getting money into the system is plagued by usual fraud problems of card TX for pure digital goods
So you handle that exactly the same way as current platforms - you use a payment processor.
> - nobody wants to build a federated system; everyone wants to build a Play/Apple/Steam store where they take 30%
I have to pay 30% already. I'd rather pay directly than with my eyeballs and brain (through ads). This is a problem, but it's better to implement a solution, and then lobby for regulation requiring an open, interoperable payment protocol.
> - winner-take-all effects are strong
Sure? We have the same problem for ads and platforms currently.
> - Play store et al already exist, why not use that?
This was already answered by vlehto in a response[1] to your comment you linked above, which you conveniently left out here. I'll quote:
> Play store does not have the content I want. And it seems overly difficult to post the content I want to make. And if I want to pay for engagement with the content, that's not an option.
> Patreon is lot closer to what suits my needs. But even that is too inflexible on how to use payments and to what you should pay. So far everybody seems to be making their micro-transaction payment models "flexible" by making the amount paid flexible. But that's exactly the one thing I want to be fixed. I'd like to host my entirely own webpage and patreon just to handle the money from exactly the kind of transaction I want.
Patreon is obviously not a micropayment platform and grossly inadequate for, well, almost anything - it's run by bad people who take large cuts and screw over creators, the friction to use it is incredibly high, and the payment model (subscriptions) does not scale well and isn't fair to smaller creators.
> - Free substitute goods are just a click away
...and yet, somehow people still pay for things online. This is quite the non-argument.
> - Consumers will pirate anything no matter how cheap the original is
Piracy is obviously evil and I'm doing my best to fight it. But this isn't an argument. In fact, the commonly-touted line "piracy is a service problem" logically implies that low-friction micropayments will make piracy less prevalent, not more.
> - No real consumer demand for micropayments
See above points about the market being subsidized by ads.
> => lemma from previous 3 items: market for online goods is efficient enough to drive all marginal prices to zero
...and yet people still pay money for things.
> - existing problem of the play store letting your kid spend all the money
This has nothing to do with microtransactions at all. This is just a platform permissions problem.
> - friction: it would be great if you didn't have to repeatedly approve things, such as a micropayment for every page of a webcomic archive. But blanket approval lets bad actors drain the jar or inattentive users waste it and then feel conned
It's trivial for someone to break up their webcomic into chapters and have users pay for the bundle. If a particular comic creator doesn't, then they'll very quickly implement that as their readers get incredibly annoyed and leave. And the vast majority of comic creators will use an existing platform to host instead of rolling the microtransaction system themselves. As for being conned? We handle that in exactly the same way as current digital storefronts.
> - first most obvious model for making this work is porn
And? I don't see how this is relevant.
> - Internet has actually killed previously working micro-ish payment systems
See above points about the market being subsidized by ads, systems being launched before their time, etc.
> surviving ones like premium SMS and phone have a scammy, seedy feel
That's purely a function of those things, and is not intrinsic to microtransactions, as evidenced by F2P games.
> - accounting requirements: do you have to pay VAT on that micropayment? do you have to declare it? Is it a federal offence to sell something to an Iranian or North Korean for one cent?
We handle this in exactly the same way as current platforms/payment processors.
> You seem to have conjured the impression that micropayment systems have to be radically different than current payment models, which is wildly mistaken.
His claim is that the existing system has fraud, therefore micro-transactions will have analogous thing he named "micro-fraud" — so you agree with him now?
As a Harper's and Lapham's Quarterly subscriber, I have been a huge fan of his quirky editorial style.
Specifically, I'd like to call out his podcast ("The World In Time"). Its past episodes remain treasure troves of wisdom, with LL's resonant voice asking the kind of engaging questions that are a rarity these days. Highly recommended.
Thanks for bringing this up - while we talk about Soros' forecasts and comparing them against those of an LLM, in the end Soros is not a forecasting tool, it's an analytical framework.
There is a gap between quant modeling and geopolitical analysis that we seek to fill. Specifically, quant models are great at capturing statistical regularities in financial time series but typically treat geopolitical shocks as exogenous noise. Meanwhile, geopolitical analyses in the policy and intelligence communities (with the exception of Bueno de Mesquita [BdM]'s work) provide deep contextual reasoning but rarely produce probabilistic scenario structures or asset-level transmission mappings that can directly inform capital allocation.
We will be shortly publishing a technical preprint laying out the Soros framework in full, but the TL;DR is: we model geopolitical events (or crises in the literature) as partially observed ("fog of war") stochastic games with multiple actors jostling for control over resources. We map out actors across various axes (think of these as actor embeddings), identify key decision points, and enumerate paths across them to estimate scenario probabilities. The scenarios in turn have associated transmission flows and market implications. We will evaluate those as mentioned in the sibling comment. Happy to discuss more.
reply