Ted Merz

Ghostwriting for CEOs at Principals Media. Former Global Head of News Product at Bloomberg.

[Ted Merz @TedMerz]

@IanRountree totally

[Ted Merz @TedMerz]

@sjwhitmore I went through this with two kids and i still found this educational and enjoyable

[Ted Merz @TedMerz]

@frazerrice @wealthactually that was a really fun conversation

[Ted Merz @TedMerz]

@BrookeDukellis I need that hat

[Ted Merz @TedMerz]

@BrookeDukellis this is definitely the content we've been waiting for. can you set up a camera and auto detect and classify these trends at scale and you're ready to raise !

[Ted Merz @TedMerz]

agree with this. also endorse the growth of the word ""maxxing" to be applied to and modify more situations in life

[Alex & Books 📚 @AlexAndBooks_]

You need to be bookmaxxing. Read more books. Borrow more books. Buy more books. Recommend more books. Gift more books. Discuss more books. Review more books. Reread more books. Quote more books. Join more bookclubs. Take more notes on books. Apply more books. Do more with books.

[Ted Merz @TedMerz]

@frazerrice I feel like I'm getting closer to that every decade

[Ted Merz @TedMerz]

@LizWebsterSBF @TheresaAFallon one of the biggest own goals of all time

[Ted Merz @TedMerz]

this is true for so many metrics. Publicly tracking numbers distorts the human behaviors. Although I agree public leaderboards are bad, if the company is still conveying the message as many are that token usage will determine whether you keep your job or not .. you will still see distortions

[Ted Merz @TedMerz]

Met a guy at NY Tech Week who showed me the "burner sneakers" he wears when he travels. The brand is Jousen and they cost $21.99 on Amazon. The world is amazing!

[Ted Merz @TedMerz]

What I love about this post is that showcases some of the real-life complexitites that companies have to handle to get to the point where they can get value about of the data. There is a lot of hand waving about how you just run Claude Code and solves everything. At the enterprise/ corporate level its different.

[Ted Merz @TedMerz]

On stage in a fireside chat for NY Tech Week with Artur Kiulian, founder and CEO of Principle.

Not planned but I wore all black; he wore all white.

We talked about the challenges and limitations of using content to feed LLMs as well as the concerns about ROI and the split in how the new technology is viewed between the West and East Coasts.

It's become clear to me a) how ill suited news articles are to feeding LLMs b) how much more textual information we need to generate better insights and c) that there is a major problem with so much online information coming from content marketing, which can be factually misleading.

@futuredesignguy

[Ted Merz @TedMerz]

The View from the Office.

I met up on Zoom with Joe O'Donnell, co-founder of Canary Data, a fintech startup that builds AI-powered software to help investors analyze opportunities.

Joe spent eight years at Tiger Global, where he ran the short portfolio. He and a friend started Canary three years ago. The original idea was to build a narrow risk-mitigation tool that would track recidivist fraudsters as a source of short opportunities.

Initially the product let users screen a company for ties to bad actors. Then it expanded to review anomalies related to accounting, disclosures, insider sales and fundamentals. About a year ago, Joe realized he could build a generalized investment analyst.

This week, Canary announced series B funding from Tiger and Arena Holdings, which is run by Feroz Dewan, Joe’s former boss at Tiger. According to the Wall Street Journal, customers include Tiger and Flight Deck Capital; annual prices range from mid-five to mid-six figures.

Joe told me he’s trying to codify into software the hard-fought lessons he learned at Tiger. He said that on his first day at the firm Feroz Dewan handed him 10 investment-committee presentations. Every one was better than anything Joe had ever done. He realized the bar was high.

Joe argues that most AI fintech products aim to cut costs by saving time. Canary is focused on improving returns by sharpening judgement. He said: returns come from good judgement not the speed of looking at ideas.

The flagship product is called SuperAnalyst. It generates questions, builds research plans and generates answers using custom AI models, public data and alternative data such as credit card or web traffic data.

To generate trade ideas, Canary built an AI screening tool and has begun releasing AI idea generation agents, the first of which is named Stanley – named after the legendary hedge fund manager Stanley Druckenmiller – that focuses on cyclical longs and shorts.

Canary is delivered via APIs, Web interface and now, MCPs. The future goal is to build out custom analysts for funds so firms can import their own data and customize the “personality”, i.e. focus on longer or shorter time frames with more optimistic or skeptical outlooks.

Joe lives in California but jokes that he works West Coast, East Coast, and sometimes European hours.

[Ted Merz @TedMerz]

There’s nothing quite like New York Tech Week, which started on Monday.

It is well organized and decentralized at the same time.

Unlike other major conferences like Davos, Art Basel Miami or South by Southwest, Tech Week isn’t run by an organization that arranges events and sells tickets.

It’s coordinated by the venture firm Andreessen Horowitz which puts out a calendar of breakfasts, and panel discussions and cocktails. Anyone can apply to attend, but the best events fill quickly.

The fact the “organizer” doesn’t have to put out capital has allowed it to grow quickly over the past several years. Yesterday alone there were record 459 events on the docket.

Most of the event is set up leveraging tweets on X and the Partiful platform to keep track of attendance. Months in advance anyone can tweet at or DM a16z partner Katia Ameri. She vets them and adds them to the list.

TECH WEEK by a16z benefits from the promotion and visibility and the firm also gets the list of attendees.

I attended the a16z New Media cocktail yesterday where I met Katie Kirsch, Karina Bao, Paula Hübner Wehmeyer, Maya Agnihotri, Sonia Thosar, Jack Randall, Sid Balaga, Calvin Campos and Claire Barthelemy.

I’m speaking at an event today
https://t.co/yBWQ7r0cFL

Overall calendar
https://t.co/KV1wmTyuGZ 

@katiekirsch, @repkarinabao, @mayarox5, @JackRandalll, @sidbalaga  @KatiaAmeri

[Ted Merz @TedMerz]

@KatiaAmeri how many events are slated overall for the full week in NYC

[Ted Merz @TedMerz]

@jefielding @a16z agree. remarkable how it's grown

[Ted Merz @TedMerz]

The View from the Office.

I caught up with Howard Lindzon in Soho at La Colombe where we talked about markets, media and machines over cappuccinos.

Howard is the founder of venture firm social leverage and the online community Stocktwits. We met 15 years ago while I was at Bloomberg. We’ve continued to collaborate.

Howard’s been energized by the emergence of AI in general and Claude Code in particular. He said it is transforming his business.

AI has increased the value of data created by Stocktwits, a community of 500,000 daily active users who post about stocks, crypto and the markets.

AI can be used to identify shifts in sentiment and detect stocks the traders in Stocktwits focus on.
Howard cited several trends he's seeing. The big ones include:

--The emergence of research beyond traditional Wall Street analysts - people like Michael Burry on Substack or Citrini - suggest we are likely to see a lot more individuals writing insight.

--He expects more specialized media. There could be a YouTube channel similar to TBPN for every ticker.

--On social media, distribution matters more than follower counts, which increasingly are a vanity metric.

--The new operating system is the LLM, and distribution means being there. Forty years of plumbing built to push data to terminals gets obviated by a connector in Claude.

--Just because you can build it doesn't mean you should. Defensibility lives in brand, distribution, and customer service, not code.

--If you're not pushing code, no one notices.

--You need a sales force more than ever. Founders think a tweet replaces a quota. It doesn't. The product is cheaper to build and harder to sell.

--Algo trading is going retail. A Mac Mini and an Alpaca account can turn anyone into a one-person quant desk.

--The first transformation was information, now it's transaction. Twitter, Bloomberg, and Google rewired how we found things. LLMs plus MCP rewire how we do them.

--Reading news and buying the stock collapses into the same prompt.

@howardlindzon

[Ted Merz @TedMerz]

@skylarromines brings to mind the Buffett adage about the most important trait in a partner is "low expectations."

[Ted Merz @TedMerz]

@skylarromines this is a really good idea

[Ted Merz @TedMerz]

@Jack_Raines @HistoryBoomer definitely C