Bullpen.News From Edwin Dorsey
Get Daily Recap Feedback

Latest News

Market capOver $300M
All Over $50MOver $300MOver $2BOver $10B
All Under $1BUnder $10BUnder $100BUnder $500B
SectorsAll
Filter by sector
Monday, October 19, 2026· 9422 stock ideas· 219 writers·
19,823 views· 9,155 clicks

From X

@convequity

Convequity on X

A company gets an API key from a model provider (OpenAI, Anthropic, Google) and links it in OpenCode Console. OpenCode, the coding agent, uses that key to run models at the provider, billed to the company. It is a bearer credential: anyone with a copy can spend on the account. OpenCode disclosed unauthorized access to a legacy system being decommissioned. Keys for about 1.4% of Console users who linked external providers may have been exposed. OpenCode calls it a distillation attack. Users report charges for runs they never made. Provider terms generally make the key owner responsible, so the bill can land on the customer. A copied key works at the provider until revoked. OpenCode already runs on Cloudflare Workers, and Console is a gateway with budgets. Direct calls with a stolen key bypass Console and its budgets. Whether that legacy host ran on Cloudflare is undisclosed. In our view, customers need a control layer they own that keeps the raw provider key away from the agent and its vendor: a vault-held key, or Cloudflare ($NET) AI Gateway prepaid billing with no customer provider key, plus scoped keys, spend caps, per-user limits, alerts and full logs. F5 ($FFIV), Kong and LiteLLM offer related controls. We cannot say these would have stopped this breach. They can limit blast radius, spend and detection time: metrics we think a CFO will ask for before the next rollout.
@convequity

Convequity on X

As AI agents take over more software work, a SaaS vendor's future comes down to how much of the interface, workflow and data layers it keeps. Ahead of this quarter's software earnings, we will watch how far management is willing to change products and business models to keep that work, because strong defenses today can erode if a company refuses to change. Our SaaS Survival Scorecard evaluates the structural resilience to expanding agentic work, vendor adaptiveness, and what the market is pricing in. Route 3: customers keep working in the vendor's interface, workflows and data as agents do more of the tasks. Route 2: customers move to an agent interface such as ChatGPT or Claude but still rely on the vendor's workflows and data. At the weakest end, only the data layer remains. Route 1: customers replace the vendor, using AI to build and maintain all three layers themselves. Our view: leading security platforms like $PANW have the stronger Route 3 case, since analysts investigate inside them. Datadog ($DDOG) faces more Route 2 risk as developers work inside coding agents and call on it from there. Deeper dive available at Convequity.
@kevin__mak 1 click

Kevin Mak on X

$CMPS Update on Compass Pathways (1) We're currently waiting for the company to complete its rolling NDA (New Drug Application) submission to the FDA, expected this quarter. The FDA has already started reviewing portions of the application. Reminder that Compass has been granted both rolling review and a Commissioner’s National Priority Voucher, which targets a review period of roughly 1–2 months once the complete application is filed. That timeline isn't guaranteed, but this is one of the fastest review processes available. It doesn't guarantee approval either, but it clearly shows that the FDA is prioritizing getting this drug reviewed. (2) The company presented at the STAT Summit, and the CEO said they will “go slow before we go fast.” My guess is that's being interpreted in a slightly negative way, suggesting that the initial drug launch won't necessarily get quick traction. As an investor in the company, my first reaction is that I don't want to hear that. I want to hear that revenues are going to explode out of the gate. But just because I want to hear something doesn't make it true. Taking a step back, Compass has taken a deliberate approach to developing this drug and preparing for launch. The CEO's point was that bad early experiences for patients or providers could set back both Compass and the broader industry. Getting those first experiences right matters. I think calibrating expectations and not being overly promotional is a smart choice. That could mean a more measured initial revenue ramp, and I'm fine with that if it sets them up for sustained adoption. At the same time, the company has been preparing to scale. Its strategic collaborations span more than 1,000 sites, including Neuronetics/Greenbrook and other established treatment networks. There is already substantial infrastructure to build on. I think this is the correct approach. (3) The company has now explicitly communicated expected annualized drug pricing of $50k–$65k per patient, based on an average of roughly three doses per year. That implies approximately $17k–$22k per dose, before clinic administration and monitoring costs. The final launch price hasn't been announced, but management has given us a much clearer pricing framework. In terms of M&A, which isn't part of my base case, my guess is the best window is after the label is set. At that point, a potential buyer will know much more clearly what they're buying and how it can be commercialized. From a positioning perspective, I think completing the NDA submission could bring new investors into the stock. It removes another layer of timeline uncertainty and puts an FDA decision much more directly in view. My guess is that matters to investors who have been watching but haven't wanted to get involved yet. My long-term outlook on the stock is unchanged. I think they get a good label, and I think concerns about the length of clinic visits are overblown. The market is massively underserved, and Spravato (which I believe COMP360 will offer a more attractive alternative to) is already selling at roughly a $2.5B annualized rate. My PT remains $45, and my full write-up is below. https://t.co/rVZOU3yIz6 Disclosure: Creek Drive and its affiliated funds holds a positions in CMPS. Views expressed are current opinions for information only, subject to change and are not investment advice or a recommendation to buy or sell any security. Investing involves the risk of loss and past performance is not indicative of future performance. See the disclosures link in my bio for more important information.

Investor Letters

All letters

Free Recaps of the best new stock ideas, delivered every morning.

Read past recaps

Build-A-Bear: A Narrower Path

Build-A-Bear's push into prestuffed toys and broader retail distribution tests brand stretchability, but recent results warrant caution.

BBW $317M
TSOH Investment Research 5 clicks

Do You Trust Apple?

Apple’s HomeView strategy ties Siri, privacy and smart-home control to consumer trust in its vision for the home.

AAPL $5T
Stratechery by Ben Thompson – On the business, strategy, and impact of technology. 2 clicks

Semtech Analyst Day Review

Semtech raised Q3 guidance and targets $3.2 billion of fiscal 2030 revenue, excluding upside from its AI data-center optical and copper…

SMTC $16B
CV's Substack 11 clicks
CDLR · long

Cadeler A/S

Cadeler (CDLR) is the world’s largest pure-play owner-operator of wind turbine installation vessels, listed in Oslo and on the NYSE.

CDLR $2.1B
Value Investors Club 4 clicks

The Strategic Value of eDreams

EDreams’ Prime travel subscription model makes the €3.6 billion European OTA a potential Uber, Airbnb or Amazon acquisition target.

EDDRF $545M
310 Value’s Newsletter 2 clicks

Notes and Updates

ECIP recipient banks could repurchase Treasury preferred shares at steep discounts, potentially lifting book value, while Hawaiian Electric…

HE $1.5B
Exploring with Alluvial Capital

JBG Smith

JBG Smith trades at an estimated 40% discount to NAV, backed by National Landing multifamily assets, Amazon HQ demand and $1.3 billion of…

JBGS $813M
Raging Bull Investments

Have a source we’re missing? Send your pick

Recent recaps

Latest recap

Sunday, October 18 32 stock ideas
  • What Lies Beneath: The Developer Building a Data Center on a Former Army Dump Laced With Munitions and Carcinogens Hunterbrook
  • DigitalOcean CEO: ‘We are not going to be extraordinarily leveraged like some others’ The Technology Letter
Read the recap

All recaps

Publications we follow

Every publication on this site, hand-picked.

130 publications

Newsletters 108

YouTube 4

Blogs 17

Seeking Alpha 52 of 239

Nominate a publication

X accounts we follow

Every account on this site, hand-picked.

89 accounts
Nominate an account

Investor letters we follow

189 firms

Sectors

Ordered by how often our editor reaches for them.

Anything we’re missing?

Send us a pick