A lab for experiments at the edge of finance + Weekly segments on prediction markets & crypto.

Internet
Our Deep Research AI has been on a tear the past couple weeks. 700,000+ queries since Oct 2025 Powering 13 partner protocols, running 24/7 TODAY: We're opening it up for anyone to use, visit factsai.org and get set up. • Create multiple API keys (teams, agents, products) • Track historical queries/usage per key • Flat pricing: $0.01 per query In the next few weeks, we'll integrate more protocols improving the model as we go 🧠
16
26
97
11,826
Polyfactual retweeted
Derive just had its biggest notional week ever. And onchain options are just getting started.
19
37
337
27,821
Polyfactual retweeted
Bitcoin closed at $75,600 on Sep 15. Six days later it was $11,000 higher It hasn't come within 9% of that price since For anyone who held through a crash of more than 50%, a 7% drop can feel like the start of the next one. Selling feels safe bitcoin:native fell 7% from Sep 3 to Sep 15, then climbed for six days without a real pullback Anyone who sold at $75,600 needs a 10% drop to buy back at the same price. On Polymarket, that's a long shot before October. With four days left: >Back above $85,000: 72% >Dip to $80,000: 15% >Dip to $75,000: 1.5% So the seller can buy back more than 10% higher, or keep waiting. Neither is easy. Sometimes that fear is right. Bitcoin rose about 30% from its February low to its May high, then fell below that low in June. Where were you on the September dip?
10
4
22
1,279
Zoning in further on the agentic trading bit, This year, Robinhood took a step to upgrading the experience with Brokerage APIs to fit the MCP structure of Ai agents. What this basically means is, They implemented a pathway for agentic trading on the RH platform that lets you connect a third-party agent like Claude, ChatGPT, Grok and other MCP-compatible tools to a dedicated Robinhood account with a budget and then let that agent place trades across equity, options and crypto orders. By the end of Q2, nearly 100k customers had opened these accounts, holding more than $100 million in custody. The interface is no longer only screens built for people but also a programmable endpoint software can call Would be interesting for sure to hear what Vlad has to say this week at the Summit on this regard
Robinhood’s third annual HOOD Summit, themed “Engines of Creation” begins on Tuesday, September 29th, where @vladtenev will deliver a keynote and unveils new products aimed at active users of the Robinhood eco Day 2 follows on Wednesday, September 30, with another full day of sessions from Robinhood executives and industry guests including Cathie Wood, David Hoffman of Bankless, and Tom Sosnoff. Both days will be livestreamed on the Robinhood app, YouTube and X. The timing of this particular summit matters because the company is coming into this event after a stretch of product and positive regulatory news ranging across Robinhood Chain, Tokenized assets available in more than 120 countries, Agentic Trading for AI-connected accounts and the SEC’s Innovation Exemption for tokenized stocks. The Summit is where Robinhood is expected to show what that stack means for everyday traders, rather than only for the brokerages
1
11
975
Robinhood’s third annual HOOD Summit, themed “Engines of Creation” begins on Tuesday, September 29th, where @vladtenev will deliver a keynote and unveils new products aimed at active users of the Robinhood eco Day 2 follows on Wednesday, September 30, with another full day of sessions from Robinhood executives and industry guests including Cathie Wood, David Hoffman of Bankless, and Tom Sosnoff. Both days will be livestreamed on the Robinhood app, YouTube and X. The timing of this particular summit matters because the company is coming into this event after a stretch of product and positive regulatory news ranging across Robinhood Chain, Tokenized assets available in more than 120 countries, Agentic Trading for AI-connected accounts and the SEC’s Innovation Exemption for tokenized stocks. The Summit is where Robinhood is expected to show what that stack means for everyday traders, rather than only for the brokerages
6
1
10
1,467
Bitcoin closed at $75,600 on Sep 15. Six days later it was $11,000 higher It hasn't come within 9% of that price since For anyone who held through a crash of more than 50%, a 7% drop can feel like the start of the next one. Selling feels safe bitcoin:native fell 7% from Sep 3 to Sep 15, then climbed for six days without a real pullback Anyone who sold at $75,600 needs a 10% drop to buy back at the same price. On Polymarket, that's a long shot before October. With four days left: >Back above $85,000: 72% >Dip to $80,000: 15% >Dip to $75,000: 1.5% So the seller can buy back more than 10% higher, or keep waiting. Neither is easy. Sometimes that fear is right. Bitcoin rose about 30% from its February low to its May high, then fell below that low in June. Where were you on the September dip?
10
4
22
1,279
Mortgage rates at 7.03%, 10-year yields at 5.18% Polymarket's odds of an October Fed hike jumped from 50% to 66.5% in three days The 10-year was at 5.01% on Sep 16. That day, Warsh said he would be "hard pressed to describe broad financial conditions as restrictive" On Sep 23, S&P Global's flash PMI for U.S. business activity hit 58.4, its highest in over five years The same day, Governor Barr gave a speech on housing costs. He said that in his base case "further policy adjustments are likely to be needed" Polymarket now: >October hike 66.5% >December hike 71.5% >At least one more hike this year 91.5% In their September projections, only 4 of 18 Fed officials expected two more hikes this year Next: PCE Sep 30, payrolls Oct 2, CPI Oct 14, Fed decision Oct 28 Is the bond market doing the Fed's job, or pushing it to hike?
7
2
14
896
Polyfactual retweeted
Kicking off our weekly show with @polyfactual Generating odds in real time with @prophetmarketai
Market Talk: Meta muse, Prediction wash trading, Crypto options! nitter.net/i/broadcasts/1qxvveXAY…
2
4
414
Payment rails on Arc Chain As we discussed earlier this week on Arc and its native Circle/USDC functionality, it only makes sense that features and products that aid payment and payment networks through stables are actively being developed on the chain, as that would most probably serve as its biggest moat For example, Arc has StableFX, an onchain FX engine for swapping between different stables such as USDC to EURC with RFQ-like quotes and settlement. And then CPN, which is short for Circle Payments Network, an institutional network for cross-border payouts between banks, service providers, etc with as the foundational layer. Speaking of @circle, they also integrated Circle Gateway, which allows for one unified USDC balance that could be spent across chains. For people or Institutions like banks with a lot of lot of Liq, USYC serves useful purposes. It's a tokenized money-market fund that works as yield bearing collateral for treasure and settlement applications Currencies you're eligible to swap with as Stables range across, AUDF ⇨ Australian dollar BRLA ⇨ Brazilian real JPYC ⇨ Japanese yen KRW1 ⇨ Korea won MXNB ⇨ Mexican peso PHPC ⇨ Philippine peso QCAD ⇨ Canadian dollar ZARU ⇨ South African rand Definitely interesting to see how all of this plays out especially with the current aggressive growth of the digitization of financial assets and the need to transact with digital/onchain tokens in local and foreign currencies
Highlighting interesting Projects on Arc Chain. There seems to be some asymmetrical upside with Arc that most people are currently overlooking due to the recent fud that came around the chain Zooming out, there are parameters that could later translate into bullish sentiment hanging around the chain. For example, Arc currently utilizes a Proof of Authority (PoA) for the security of the network which basically means the network tokens of 10 billion ARC have been minted but none of it is publicly floating but they estimate a move to a PoS (proof-of-stake) network by 2027, which would lead to the inevitable distribution of these tokens. Anyways, some projects worth looking at on Arc range across, @TollyLabs ⇨ Closest substitute to Pump Fun on Solana as a leading launch pad product. Buyback and burns are executed as revenue is generated, there are direct-to-locked-USDC pool launches so no bonding curve lag. Fees are also split with creators and it’s currently around a $5m mcap with some decent room to run. @Morpho ⇨ The strongest independent credit layer on Arc with $227m in TVL already. They’re building multiple credit functionality on Arc with Circle helping with the distribution across the eco. $MORPHO is currently $2.58, with an Ath of $4.17 in Jan 2025. @Bullcheese_fun ⇨ This is also a launchpad but they have an interesting application of launching memes. They don’t utilize bonding curves - you launch a token and it goes straight to a Dex on Arc and is powered by TrustSwap. Creators get paid 0.75% of volume and potentially 75% of fees. Their goal is to “limit hurdles” Creators face when trying to push a token. They don’t have a token yet but creators can in $SWAP, which is around the $4m mcap mark. syrupUSDC by @maplefinance ⇨ This is Maple’s institutional credit wrapper on Arc, serving as a a plug and play earn asset for Neobanks who don’t want to build a lending desk. They’re currently offering 4.2% monthly APY and 6.7% weekly APY on syrupUSDC @Solonlabs1 ⇨ They primarily aid Lending tokenized assets operations by letting you use verified stocks/equities as collateral in exchange for USDG, which then allows you to supply this USDG into a Morpho vault and earn on it. They’re also building their launchpad on Arc and currently sitting on a market cap of $240k . Still definitely very early days for arc but interesting to watch what happens over the coming months, rest assured we’d keep you updated!
3
2
12
960
Polyfactual retweeted
Democrats need to net four Senate seats. Polymarket has them favored in five Republicans still have a 34.5% chance Three look strong: North Carolina at 94.5%, Alaska at 78.5% and Maine at 73.5% The fourth seat is the harder part. The other two, Ohio and Texas, are at 62.5% and 61.5%. Iowa is close behind at 48.5%. In Texas, two new polls have Talarico ahead. Marist has him up 6 among registered voters. TSU has him up 1 among likely voters Democrats also have to hold their own seats. The shakiest is Michigan at 74.5%, then New Hampshire at 85.5%. A week ago, Democrats were at 58.5%. Now they're at 65.5%. The three races that moved the most this week: ➤ Michigan: 61.5% → 74.5%. Suffolk has El-Sayed up 47-40 ➤ Alaska: 65.5% → 78.5%. AARP has Peltola up 53-47 in the final ranked-choice round ➤ Iowa: 40.5% → 48.5%. Marist has Turek up 50-42 among registered voters Reuters/Ipsos also put Trump's approval at 32%, his lowest in their polling Since 1946, the party in the White House has lost four or more Senate seats in 10 of 20 midterms. That's history, not a forecast Which one gives Democrats the fourth seat: Ohio, Texas or Iowa?
10
4
17
1,374
Market Talk: Meta muse, Prediction wash trading, Crypto options! nitter.net/i/broadcasts/1qxvveXAY…
1
1
10
1,073
No surprises to see Hyperliquid's ecosystem win the poll, especially with Hype breaking all time highs and more projects experimenting with HyperEVM. Robinhood's eco came in close at second which again, is by no surprise as everyone seems to be trying to get exposure to tokenized assets, a relatively new vertical which is starting to really take off
You had to bet on one ecosystem for the next 12 months, which are you choosing?
3
1
11
1,011
Democrats need to net four Senate seats. Polymarket has them favored in five Republicans still have a 34.5% chance Three look strong: North Carolina at 94.5%, Alaska at 78.5% and Maine at 73.5% The fourth seat is the harder part. The other two, Ohio and Texas, are at 62.5% and 61.5%. Iowa is close behind at 48.5%. In Texas, two new polls have Talarico ahead. Marist has him up 6 among registered voters. TSU has him up 1 among likely voters Democrats also have to hold their own seats. The shakiest is Michigan at 74.5%, then New Hampshire at 85.5%. A week ago, Democrats were at 58.5%. Now they're at 65.5%. The three races that moved the most this week: ➤ Michigan: 61.5% → 74.5%. Suffolk has El-Sayed up 47-40 ➤ Alaska: 65.5% → 78.5%. AARP has Peltola up 53-47 in the final ranked-choice round ➤ Iowa: 40.5% → 48.5%. Marist has Turek up 50-42 among registered voters Reuters/Ipsos also put Trump's approval at 32%, his lowest in their polling Since 1946, the party in the White House has lost four or more Senate seats in 10 of 20 midterms. That's history, not a forecast Which one gives Democrats the fourth seat: Ohio, Texas or Iowa?
10
4
17
1,374
Highlighting interesting Projects on Arc Chain. There seems to be some asymmetrical upside with Arc that most people are currently overlooking due to the recent fud that came around the chain Zooming out, there are parameters that could later translate into bullish sentiment hanging around the chain. For example, Arc currently utilizes a Proof of Authority (PoA) for the security of the network which basically means the network tokens of 10 billion ARC have been minted but none of it is publicly floating but they estimate a move to a PoS (proof-of-stake) network by 2027, which would lead to the inevitable distribution of these tokens. Anyways, some projects worth looking at on Arc range across, @TollyLabs ⇨ Closest substitute to Pump Fun on Solana as a leading launch pad product. Buyback and burns are executed as revenue is generated, there are direct-to-locked-USDC pool launches so no bonding curve lag. Fees are also split with creators and it’s currently around a $5m mcap with some decent room to run. @Morpho ⇨ The strongest independent credit layer on Arc with $227m in TVL already. They’re building multiple credit functionality on Arc with Circle helping with the distribution across the eco. $MORPHO is currently $2.58, with an Ath of $4.17 in Jan 2025. @Bullcheese_fun ⇨ This is also a launchpad but they have an interesting application of launching memes. They don’t utilize bonding curves - you launch a token and it goes straight to a Dex on Arc and is powered by TrustSwap. Creators get paid 0.75% of volume and potentially 75% of fees. Their goal is to “limit hurdles” Creators face when trying to push a token. They don’t have a token yet but creators can in $SWAP, which is around the $4m mcap mark. syrupUSDC by @maplefinance ⇨ This is Maple’s institutional credit wrapper on Arc, serving as a a plug and play earn asset for Neobanks who don’t want to build a lending desk. They’re currently offering 4.2% monthly APY and 6.7% weekly APY on syrupUSDC @Solonlabs1 ⇨ They primarily aid Lending tokenized assets operations by letting you use verified stocks/equities as collateral in exchange for USDG, which then allows you to supply this USDG into a Morpho vault and earn on it. They’re also building their launchpad on Arc and currently sitting on a market cap of $240k . Still definitely very early days for arc but interesting to watch what happens over the coming months, rest assured we’d keep you updated!
3
3
16
1,572
You had to bet on one ecosystem for the next 12 months, which are you choosing?
33% Robinhood
38% Hyperliquid
29% Solana
21 votes • Final results
4
1
7
1,460
Polyfactual retweeted
Day 16✅ been digging deep into @Polyfactual and there’s more to this than just an AI tool for @Polymarket prediction markets have a lot of information scattered across news, social media, market data, onchain data and different sources Polyfactual brings this information together and uses AI to break down markets, sentiment, risks and other factors that could affect the outcome the trader value is simple instead of spending time going through multiple sources before taking a position, you can use Polyfactual to speed up the research process and get a clearer picture of what’s happening around a market this becomes even more useful when you’re tracking multiple markets or when new information suddenly changes the market they also have a chrome extension, so you can access the analysis directly while using Polymarket but what really caught my attention is FactsAI they’re building an API that other prediction market apps can integrate for research and intelligence they also have FactsBase for OSINT and intelligence, alongside a trading terminal and other products so they’re not just building another Polymarket dashboard they’re trying to build infrastructure around prediction market research and trading research → intelligence → analysis → trading the main thing i’ll be watching is how accurate and useful the AI analysis proves to be over time because the real value isn’t just getting an AI-generated confidence score it’s whether the information actually helps traders understand markets faster and make more informed decisions definitely a tool i’ll be testing more
Day 15 ✅ been looking at @DataDashxyz and what they’re building around prediction markets @datadashxyz is making it easier to find and study markets on @Polymarket without having to manually go through everything yourself you can see things like market prices, 24h price movement, trading volume, number of traders, time left before the market resolves but the interesting part for me is that it goes beyond just showing Polymarket markets you can also track trader activity and look at things like smart money, wallet behaviour, signals and cohorts so instead of just asking “what is the market saying” you can also start asking “who is trading this market” “what are they doing” “how have they performed before” “what other markets are they trading” that can make research a lot easier for example if i see a Polymarket market trading at $0.65 DataDash can help me look beyond that 65¢ and understand the activity behind the market how much volume is coming in how many people are trading it how the price has moved and which wallets are active around it i think the smart money and wallet tracking side is one of the more interesting parts because you can follow traders instead of just looking at the market itself it doesn’t mean every wallet you see is automatically right tho the data is there to help with research, you still have to do your own work before taking a position overall i see DataDash as a research layer sitting on top of prediction markets making it easier to discover markets, study traders and understand what is happening behind the numbers
9
6
27
885
Polyfactual retweeted
The White House demanded rate cuts. The hike is at 87.5¢ Fed decides today, 2pm ET. 25bp hike 87.5¢, no change 11.5¢ on Polymarket Sep 4, Trump on Truth Social: "lower the interest rates because the U.S.A. is a much stronger credit than it was just a short time ago!" neither that post nor Vance the day before moved the price. the repricing was data Warsh has held rates twice since taking over in May. in July the vote was 9-3, with Hammack, Kashkari and Logan all wanting a hike. that was before either hot print > the second hike is priced in December - October hike 35.5¢ - December hike 58.5¢ Polymarket puts 50bp of hikes for the whole year at 44.5%. the month markets put the second 25 in December October meets six days before the midterms. December brings new projections, October does not. the calendar explains that spread better than the data does Warsh put PCE at 3.7% over twelve months and 4.1% over six at Jackson Hole. the short window is running hotter, and a second hike is already in the price > the dots will not settle it at 2pm they give one median for the whole year. they cannot name a meeting. Warsh did not submit one in June: "it's not helpful in the conduct of policy" the open question is whether the second hike lands six days before an election or a month after it Data or Calendar. Which one is October pricing?
The Fed almost never hikes just once In the past 30 years there has been only one exception: 1997 After CPI, Polymarket is heavily pricing in +25 bps on Wednesday - around 80% In March 1997 the Fed hiked from 5.25% to 5.50% and stopped. The Asian crisis kept them from going further. Cuts only started in 1998, after Russia and LTCM The other cycles were series: >1994-95 >1999-2000 >2004-06: 17 consecutive hikes >2015-18: 9 hikes >2022-23: 11 hikes A single +0.25 does almost nothing on its own. Once the Fed moves up, the market immediately starts pricing in the next hike Midterms are 7 weeks away Trump didn’t pick Warsh for a hiking series, and the White House is no longer hiding that Wednesday is less about the +25 bps than the new dot plot - and whether Warsh signals "we keep going" or "this is not a cycle" They may skip October because of the election. December is still on the table On Polymarket, one hike in 2026 is 43%, two hikes 41% An every-meeting series may not happen A single hike for the year is no longer the obvious base case
6
1
14
3,890
Polyfactual retweeted
The Fed could hike again six days before the midterms On Wednesday morning that was a 35.5¢ on Polymarket. It is 56¢ now Asked by the New York Times whether this hike starts a sequence, Warsh said he is "not in the forward guidance business" and would "not pre-judge any future decisions we make" Asked by Politico about the market's focus on the August CPI: "Data point dependence is a dangerous preoccupation" The vote was unanimous. Warsh, as in June, submitted no projection of his own. Of the 18 who did, 16 see at least one more hike this year. Not one of those dots names a meeting Polymarket: > October hike 56¢ > December hike 68¢ Those add to 124¢, which only works if a hike at both meetings is live The banks split the same way > After the meeting Goldman tore up its one-hike-for-the-year call and moved to October > Morgan Stanley had gone in calling September and December > Bank of America's Aditya Bhave, also before the vote, put market pricing at "close to 100bp of hikes in total, over the next year or so" and expected less than that, all of it this year The October meeting is Oct 27-28, with no new projections. December has them. That calendar has not changed since Wednesday. The gap between the two months was about 25 points that morning. It is 12 now How much of that 12-point gap is the election?
The White House demanded rate cuts. The hike is at 87.5¢ Fed decides today, 2pm ET. 25bp hike 87.5¢, no change 11.5¢ on Polymarket Sep 4, Trump on Truth Social: "lower the interest rates because the U.S.A. is a much stronger credit than it was just a short time ago!" neither that post nor Vance the day before moved the price. the repricing was data Warsh has held rates twice since taking over in May. in July the vote was 9-3, with Hammack, Kashkari and Logan all wanting a hike. that was before either hot print > the second hike is priced in December - October hike 35.5¢ - December hike 58.5¢ Polymarket puts 50bp of hikes for the whole year at 44.5%. the month markets put the second 25 in December October meets six days before the midterms. December brings new projections, October does not. the calendar explains that spread better than the data does Warsh put PCE at 3.7% over twelve months and 4.1% over six at Jackson Hole. the short window is running hotter, and a second hike is already in the price > the dots will not settle it at 2pm they give one median for the whole year. they cannot name a meeting. Warsh did not submit one in June: "it's not helpful in the conduct of policy" the open question is whether the second hike lands six days before an election or a month after it Data or Calendar. Which one is October pricing?
8
4
21
1,647