what are you selling today?

[gfp]

Who is buying my stock I wonder...  S-3 filed on Friday night for an ATM stock offering of up to $500m.  And yet someone keeps gobbling up my A-shares.  Is it a robot I wonder?

 

Edited January 20 by gfp

[UK]

59 minutes ago, gfp said:

company is Tulips (their operating subsidiary is a company called Bloomia).  Tulip-mania!

Keep it, maybe this is a place where all these usd reserves will be soon rotated:)))

[Saluki]

I got assigned some shares from puts I sold a while back on Smith and Wesson and XIFR (formerly Nextera Enera Partners, not XPLR), and OXY. I made money on SWBI and XIFR, but lost money on the OXY shares. 

[gfp]

On 1/20/2026 at 9:27 AM, gfp said:

Who is buying my stock I wonder...  S-3 filed on Friday night for an ATM stock offering of up to $500m.  And yet someone keeps gobbling up my A-shares.  Is it a robot I wonder?

 

 

I feel like this is my only job at this point

 

 

Edited January 21 by gfp

[cubsfan]

2 hours ago, gfp said:

 

I feel like this is my only job at this point

 

 

 

Looks like a great job at that!

[backtothebeach]

3 hours ago, gfp said:

 

I feel like this is my only job at this point

 

 

Do you have a special technique selling such an illiquid stock? Hidden orders?

[gfp]

20 minutes ago, backtothebeach said:

Do you have a special technique selling such an illiquid stock? Hidden orders?

 

It used to be really difficult to buy and sell this stock, especially the A shares.  I would use odd-lots that didn't move the bid (an even lot on this stock used to be 10 shares but that seems to have changed around the time they joined NYSE Texas).

 

Recently I have used IB's mid price algorithm with a limit - but even that is not needed with these stupid buyers.  I just use limit orders over and over this week.  4 shares at a time.  Mostly A shares because they are trading at a large premium to the B-shares.

 

I've sold a ton of stock into this move with no liquidity issues whatsoever - as long as I am patient with my 4 share lots and set limit orders above the price.

 

I'm starting to enjoy it!  Ask me again at tax time next year

 

4 shares at a time!

Edited January 21 by gfp

[Saluki]

Sold a few ATM puts on CPNG for September. With the premium it would give me a $17.60 effective price if I'm assigned the shares, or free money if I'm not. I have some puts expiring 1/30 which will get assigned to me if the price doesn't move above $22. Since I have higher cost shares in CPNG, I think I'll hold them for 30 days, sell the costlier ones, take the tax loss to offset gains, and still keep my midsize position.  

 

Trimmed most of my VG after the pop, but still have some shares in my retirement account. I think it's still got room to move up, but I don't like having leverage in my main account when I'm semi-retired so I'd rather have dry powder if the market corrects like it did this week. 

 

 

[backtothebeach]

2 hours ago, gfp said:

I would use odd-lots that didn't move the bid

That’s what works best for me with FFH, too.

[gfp]

22 hours ago, gfp said:

 

It used to be really difficult to buy and sell this stock, especially the A shares.  I would use odd-lots that didn't move the bid (an even lot on this stock used to be 10 shares but that seems to have changed around the time they joined NYSE Texas).

 

Recently I have used IB's mid price algorithm with a limit - but even that is not needed with these stupid buyers.  I just use limit orders over and over this week.  4 shares at a time.  Mostly A shares because they are trading at a large premium to the B-shares.

 

I've sold a ton of stock into this move with no liquidity issues whatsoever - as long as I am patient with my 4 share lots and set limit orders above the price.

 

I'm starting to enjoy it!  Ask me again at tax time next year

 

4 shares at a time!

 

 

Like a broken record, I am, again, selling Biglari Holdings class A!  Come and get 'em crypto bros!  Or robots?  

 

[Spekulatius]

reducing $RTO and $ABEV a bit

[lnofeisone]

I shorted UNG calls and recycled that into Fairfax. 

[gfp]

It's getting harder to sell BH.A ... what does it mean?  Are the robots still interested in buying stock?

 

 

Edited January 23 by gfp

[Saluki]

I trimmed a little Kraken Robotics. It's kind of pricey now (like a lot of defense stocks) and I own the shares in my brokerage where I trade microcaps that my main brokerage doesn't list, so I decided to free up some cash to buy a couple of smaller names. 

[gfp]

On 1/23/2026 at 11:46 AM, gfp said:

It's getting harder to sell BH.A ... what does it mean?  Are the robots still interested in buying stock?

 

 

 

 

Is it over?  Only one fill!  It's over!

 

[brobro777]

Shorted tiny tiny position in silver futures

 

I know I'm gonna get smoked, get my ass kicked bad with this but like moth to flame, I couldn't help myself

 

 

 

 

[sholland]

Sold all my Chapters Group.  Still trading at a cheery consensus while Constellation Software and Topicus have had large draw downs and both now roughly half the P/S multiple of Chapters.

[Saluki]

Kraken Robotics had a nice day again, so I sold another 1000 shares to keep it at the weighting that I want. I really HATE paying taxes, but all of it is long term capital gains now, and I won't have to pay tax on it until 2027, so why not? 

[jfan]

Selling 15% of my cameco shares after a nice run up over the past five years. 

[formthirteen]

Throwing babies out of my portfolio. Including FDS:
 

Spoiler

Following the "Walled Garden" thesis—where the vegetable (proprietary data) is rare and the chef (AI model) is a commodity—I have identified the public companies that own the world's most valuable "vegetables."

These companies are not "tech" stocks in the traditional sense; they are data monopolies that have spent decades acquiring information that cannot be scraped from the open internet.

Here is your "Walled Garden" Public Equity Watch List.

---

Sector 1: Legal & Professional Services

The Moat: Decades of case law, tax codes, and scientific journals that exist behind paywalls.

1. Thomson Reuters (TRI)

• The "Vegetable" (Data Moat): Westlaw. This is the definitive record of the US legal system. It contains over 100 years of case law, statutes, and proprietary legal analysis that does not exist on the open web.

• The "Meal" (AI Unlock): CoCounsel.

  • Instead of just searching for a case (raw data), a lawyer now asks CoCounsel to "Draft a deposition strategy based on Delaware Chancery Court rulings from 2018-2023."

  • The "Richer/Lazier" Metric: It is projected to save legal professionals 240 hours annually (approx. $19k in billable time per lawyer). They are moving from selling data access to selling the work of a junior associate.

2. Wolters Kluwer (WKL)

• The "Vegetable" (Data Moat): UpToDate. This is the clinical decision support tool used by 2 million+ clinicians. It is evidence-based medicine verified by 7,000+ expert physicians, not scraped from Reddit or WebMD.

• The "Meal" (AI Unlock): AI Labs / UpToDate AI.

  • Doctors don't want a "chat" that hallucinates; they want a diagnostic recommendation cited with peer-reviewed certainty. Wolters Kluwer is integrating generative AI to answer complex clinical questions with zero-hallucination constraints because they own the source text.

3. RELX (RELX)

• The "Vegetable" (Data Moat): Elsevier (Scientific) and LexisNexis (Legal). They own a massive percentage of the world's scientific research and legal records.

• The "Meal" (AI Unlock): They have moved from "publishing" to "decision tools." In insurance, they use proprietary contributory data (claims history) to predict risk better than any generic model could.

---

Sector 2: Financial Intelligence

The Moat: Private market data, credit ratings, and supply chain mapping that isn't in a 10-K.

4. S&P Global (SPGI)

• The "Vegetable" (Data Moat): Capital IQ is standard, but their recent $1.8B acquisition of "With Intelligence" is the alpha. This gives them rare data on private markets (Hedge Funds, Private Credit, PE) which are notoriously opaque and invisible to ChatGPT.

• The "Meal" (AI Unlock): Private markets are the hardest to analyze because data is unstructured (PDFs, emails). S&P can now train models on this private repository to give investors an "AI Analyst" for private equity—something no public LLM can ever do.

5. Moody's (MCO)

• The "Vegetable" (Data Moat): Credit ratings and the "Orbis" database (hundreds of millions of private companies).

• The "Meal" (AI Unlock): "Research Assistant."

  • Analysts spend 40% of their time spreading financials and reading boring credit agreements. Moody's AI tool reads these proprietary documents and writes the "Credit Memo" for them.

  • The value: They effectively sell a "Junior Credit Analyst" in a box to banks.

6. FactSet (FDS)

• The "Vegetable" (Data Moat): Deep consensus estimates, supply chain relationships, and ownership data.

• The "Meal" (AI Unlock): "Fluent in Finance" (Conversational API). They are allowing developers to plug LLMs directly into FactSet's clean data stream. This prevents the "garbage in, garbage out" problem that plagues financial AI.

---

Sector 3: Real Estate & Insurance

The Moat: Physical world data (building specs, claims history) that requires boots on the ground to collect.

7. CoStar Group (CSGP)

• The "Vegetable" (Data Moat): They are the "Bloomberg of Real Estate." They have spent $5 billion+ over 30 years sending researchers to physically photograph and measure commercial buildings. You cannot scrape this data because it exists in their database, not on the web.

• The "Meal" (AI Unlock): Real estate is opaque. CoStar is using AI to predict rent rolls, vacancy rates, and market trends with precision that Zillow or a generic AI cannot match. They are also aggressively expanding into residential with Homes.com, using their data advantage to disrupt the "Zestimate."

8. Verisk Analytics (VRSK)

• The "Vegetable" (Data Moat): 30 Petabytes of insurance data. They know the claims history of almost every property in the US.

• The "Meal" (AI Unlock): XactAI.

  • When a house burns down, Verisk's software is used to estimate the repair cost. AI can now look at a photo of a damaged kitchen and instantly generate a line-item claim estimate (drywall, paint, labor) based on local pricing.

  • Labor Replacement: This replaces the tedious work of the claims adjuster.

---

The "Walled Garden" Basket Strategy

If I were constructing a portfolio based on the video's thesis, I would group these into a "Proprietary Data" Basket.

• Aggressive Growth Play: CoStar (CSGP) and S&P Global (SPGI). They are actively acquiring new "vegetables" (Private Markets, Residential) to feed their AI.

• Defensive Moat Play: Thomson Reuters (TRI) and Verisk (VRSK). Their data is deeply embedded in the workflow; removing them is like ripping out the nervous system of a law firm or insurance carrier.

 

Edited January 29 by formthirteen
Planning on buying back FDS at some point.

[Saluki]

1 hour ago, formthirteen said:

Throwing babies out of my portfolio. Including FDS:
 

  Hide contents

Following the "Walled Garden" thesis—where the vegetable (proprietary data) is rare and the chef (AI model) is a commodity—I have identified the public companies that own the world's most valuable "vegetables."

These companies are not "tech" stocks in the traditional sense; they are data monopolies that have spent decades acquiring information that cannot be scraped from the open internet.

Here is your "Walled Garden" Public Equity Watch List.

---

Sector 1: Legal & Professional Services

The Moat: Decades of case law, tax codes, and scientific journals that exist behind paywalls.

1. Thomson Reuters (TRI)

• The "Vegetable" (Data Moat): Westlaw. This is the definitive record of the US legal system. It contains over 100 years of case law, statutes, and proprietary legal analysis that does not exist on the open web.

• The "Meal" (AI Unlock): CoCounsel.

  • Instead of just searching for a case (raw data), a lawyer now asks CoCounsel to "Draft a deposition strategy based on Delaware Chancery Court rulings from 2018-2023."

  • The "Richer/Lazier" Metric: It is projected to save legal professionals 240 hours annually (approx. $19k in billable time per lawyer). They are moving from selling data access to selling the work of a junior associate.

2. Wolters Kluwer (WKL)

• The "Vegetable" (Data Moat): UpToDate. This is the clinical decision support tool used by 2 million+ clinicians. It is evidence-based medicine verified by 7,000+ expert physicians, not scraped from Reddit or WebMD.

• The "Meal" (AI Unlock): AI Labs / UpToDate AI.

  • Doctors don't want a "chat" that hallucinates; they want a diagnostic recommendation cited with peer-reviewed certainty. Wolters Kluwer is integrating generative AI to answer complex clinical questions with zero-hallucination constraints because they own the source text.

3. RELX (RELX)

• The "Vegetable" (Data Moat): Elsevier (Scientific) and LexisNexis (Legal). They own a massive percentage of the world's scientific research and legal records.

• The "Meal" (AI Unlock): They have moved from "publishing" to "decision tools." In insurance, they use proprietary contributory data (claims history) to predict risk better than any generic model could.

---

Sector 2: Financial Intelligence

The Moat: Private market data, credit ratings, and supply chain mapping that isn't in a 10-K.

4. S&P Global (SPGI)

• The "Vegetable" (Data Moat): Capital IQ is standard, but their recent $1.8B acquisition of "With Intelligence" is the alpha. This gives them rare data on private markets (Hedge Funds, Private Credit, PE) which are notoriously opaque and invisible to ChatGPT.

• The "Meal" (AI Unlock): Private markets are the hardest to analyze because data is unstructured (PDFs, emails). S&P can now train models on this private repository to give investors an "AI Analyst" for private equity—something no public LLM can ever do.

5. Moody's (MCO)

• The "Vegetable" (Data Moat): Credit ratings and the "Orbis" database (hundreds of millions of private companies).

• The "Meal" (AI Unlock): "Research Assistant."

  • Analysts spend 40% of their time spreading financials and reading boring credit agreements. Moody's AI tool reads these proprietary documents and writes the "Credit Memo" for them.

  • The value: They effectively sell a "Junior Credit Analyst" in a box to banks.

6. FactSet (FDS)

• The "Vegetable" (Data Moat): Deep consensus estimates, supply chain relationships, and ownership data.

• The "Meal" (AI Unlock): "Fluent in Finance" (Conversational API). They are allowing developers to plug LLMs directly into FactSet's clean data stream. This prevents the "garbage in, garbage out" problem that plagues financial AI.

---

Sector 3: Real Estate & Insurance

The Moat: Physical world data (building specs, claims history) that requires boots on the ground to collect.

7. CoStar Group (CSGP)

• The "Vegetable" (Data Moat): They are the "Bloomberg of Real Estate." They have spent $5 billion+ over 30 years sending researchers to physically photograph and measure commercial buildings. You cannot scrape this data because it exists in their database, not on the web.

• The "Meal" (AI Unlock): Real estate is opaque. CoStar is using AI to predict rent rolls, vacancy rates, and market trends with precision that Zillow or a generic AI cannot match. They are also aggressively expanding into residential with Homes.com, using their data advantage to disrupt the "Zestimate."

8. Verisk Analytics (VRSK)

• The "Vegetable" (Data Moat): 30 Petabytes of insurance data. They know the claims history of almost every property in the US.

• The "Meal" (AI Unlock): XactAI.

  • When a house burns down, Verisk's software is used to estimate the repair cost. AI can now look at a photo of a damaged kitchen and instantly generate a line-item claim estimate (drywall, paint, labor) based on local pricing.

  • Labor Replacement: This replaces the tedious work of the claims adjuster.

---

The "Walled Garden" Basket Strategy

If I were constructing a portfolio based on the video's thesis, I would group these into a "Proprietary Data" Basket.

• Aggressive Growth Play: CoStar (CSGP) and S&P Global (SPGI). They are actively acquiring new "vegetables" (Private Markets, Residential) to feed their AI.

• Defensive Moat Play: Thomson Reuters (TRI) and Verisk (VRSK). Their data is deeply embedded in the workflow; removing them is like ripping out the nervous system of a law firm or insurance carrier.

 

 

 

As a lawyer, I'll chime in on Lexis/Nexis. That and WestLaw are used by everyone. You get a free account in law school and they show you how to use it, then it just continues when you are practicing. 

 

However, since I've sort-of retired, I looked into pricing. It used to be that they charged by the minute (you would do research, print it out, then read it so you don't rack up huge bills). Now it seems that they have flatter pricing and its a lot less. When I looked at one of them, for a solo practitioner it was $500/month. Still outrageous, but a lot less. 

 

There are cheaper bare bones competitors like Hein Online, that charge literally 1/10th that price and are fine if all you need is to look up cases. Lexis/Nexis and Westlaw have access to thousands of databases (law reviews, trade journals, foreign materials) and Westlaw has a keynote patent/trademark that is like a hashtag, and there are thousands of them, which let you find relevant cases easier.  So it would be hard to disrupt, but it is definitely facing lower priced competitors, and maybe AI helps them? If you can have it help you write in addition to finding cases, why not?  

 

 

[formthirteen]

On 11/6/2021 at 12:44 AM, Spekulatius said:

In Qualys case, I found a fairly profitable software company operating in a niche sector with not that much competition growing in the teens. I bought it at a below 9x P/S multiple when it temporarily sold off due to some soft quarter and the long term CEO leaving to health reasons. I felt it‘s buyable but not extremely cheap, so I just bought a starter position. I actually had an opportunity to buy it even cheaper but missed the very narrow window of opportunity so just kept my small position.

 

You can get QLYS for EV/S 6.6 today. Seems like an attractive price. AI beneficiary...

[John Hjorth]

NVO - The few shares in the total portfolio, in a tax deferred account belonging to The Lady of the House [,to buy NOVO B.CPH tomorrow for the money].

[thowed]

Sold some Lotus Bakeries (LOTB.BR).  A wonderful company, but I got overexcited & bought too high. 

 

It's up about 30% YTD from low levels, and I suspect results will be stellar, but I wanted some cash, and it felt reasonable to cut.

[Spekulatius]

On 2/2/2026 at 3:01 PM, formthirteen said:

 

You can get QLYS for EV/S 6.6 today. Seems like an attractive price. AI beneficiary...

Got a decent that out of that one. QLYS isn’t expensive and that sort of software (system of record, cybersecurity ) won’t be replaced that easily, I think I would go for something faster growing like ZS here. Cybersecurity is going to be a big issue with AI because it is much easier to create malicious agents to do all sorts of things with AI.