Every man can become the sculptor of his own mind if he sets himself the task
Gotham
Joined September 2013
- Tweets6K
- Following0
- Followers2.1K
- Likes3K
NoEdgeMonkey retweeted
Interesting new use case for integrating AI into the investment process
Google's new "Mind Maps" helps organize information from source documents
Using Cloudflare $NET as an example, Mind Maps bucketed all the information from their Analyst day into tiers of buckets
NotebookLM then summarizes all information relevant within each bubble
NoEdgeMonkey retweeted
I appreciate this energy. Get drunk. Haze your customers. Give them the hard sell.
NoEdgeMonkey retweeted
I respect OP but this is more evidence IMO that Analyst & PM roles should just be collapsed into one. Ppl should manage risk earlier in their careers. Otherwise politics/emotional nonsense interferes with generating returns when job is hard enough & industry has headwinds.
Very very fine line between gaining his respect vs turning him against you. Level of caution when toeing the line should be directly proportional to your PMs EQ, reservoir of intellectual goodwill you have accrued, how deep your conviction is, and where your IQ maps vs your PMs. And wordsmithing when “arguing” is an absolute prerequisite.
NoEdgeMonkey retweeted
Unfortunately for too long we've rewarded run-rating the most recent quarter of metrics & extrapolating mindlessly vs digging into the core of growth sustainability, unit economics, industry structure, etc. It's a HUGE negative for young investors developing a sound process.
NoEdgeMonkey retweeted
Redfin got acquired today. Shout out their CEO for the one of the wildest things I've ever heard on an earnings call
- Analyst: What if mortgage rates don't come down?
- Glenn: We'll drink our own urine or our competitors blood. Stay in the foxhole.
NoEdgeMonkey retweeted
I think China's second DeepSeek moment is here.
This AI agent called 'Manus' is going crazy viral in China right now.
Probably only a matter of time until it hits the US.
It's like Deep Research + Operator + Claude Computer combined, and it's REALLY good.
NoEdgeMonkey retweeted
$DMRC looks very interesting here. Most investors have likely given up after getting burned (again), but I think this is the most actionable it's ever been.
Here’s how I see it potentially playing out over the next 12-16 months...
NoEdgeMonkey retweeted
Just keeping a running list for myself to repost in future:
1. MSTR now represents a mid single digit of all converts in US
2. Meme coin bro buys 6.2m banana to eat
3. 10 year old rug pulls greedy meme coin people for $30k -> someone kidnaps his dog and asks for money back
4. Some of the best L/S guys I know giving up shorting stocks
5. 2/3 my feed is not ICBM rockets but “retardio” and “fartcoin” 🚀🚀
6. DJT rumor to be buying BAKKT (FT)
7. James Howell is paying 10s of millions to search landfill for lost bitcoin
8. SBF would be richest man in world right now
9. Scottie Pippen knows Satoshi and talks to him in dreams
10. TikTok “ChillGuy” meme coin is now $400m mkt cap and up 400,000%
Feel free to add so we can catalogue this time in history
NoEdgeMonkey retweeted
$CCOI reported results that are still messy, but I think we are on the cusp of the growth story cleaning up in a material way.
There is a SOTP/NAV story that is moderately interesting but mostly provides downside protection and an attractive skew, but it’s not enough value extraction to make a good equity story. Let’s talk about the growth engine: waves.
What caught me by surprise this quarter is that the wave backlog hadn’t been touched or groomed along the way at all. I can appreciate and understand why (don’t ask if you can’t install yet), but it’s a little bizarre since Cogent talks to these large wholesale customers regularly. So the 1,500 orders that dropped out of the funnel didn’t surprise me (there were over 2,300 sitting in there for over a year) - it’s that those weren’t cleaned out previously.
That was disappointing because it means there’s less visibility into the aggregate ARR sitting in the order book, waiting to be fulfilled. But it shouldn’t be surprising that many customers didn’t wait for Cogent for >12 months to provision those orders - those are unacceptable lead times that you hear as reasons for shifting share away from other major providers who are that delayed.
Gross new orders were about 1,000 in Q3 and 900 in Q4. Those started coming in when Cogent was aggressively POPing new locations (>2x in 2H) but could not yet provision. In other words, it was a healthy level of demand but still below where they should be going forward now that the network is done and ready to scale on the standardized workflow. Nonetheless, 1,000 orders at $2,000 ARPU = $24M of gross ARR added each quarter. They have field capacity to install 1,500 orders per quarter, or $36M or ARR. If there’s more demand, I’m sure they’ll find resources to increase their installation capacity.
The rubber meets the road right now. On callbacks, Cogent says they have nearly 400 orders in the 30-day installation window right now. I’m guessing that means they’ll install maybe 600-800 orders in Q1 given the ramp-up period from moving orders from the funnel into the installation queue, but I don’t know how they started the year. If they exit March with a full queue of 500/month and remain there for a while as they clear backlog and add new orders, the growth story will start to shine through by Q1 or Q2 results.
NoEdgeMonkey retweeted
I've looked through every stock in the Russell 2000 many times. It is full of:
1. Total trash companies, debt zombies, de-spacs who's equity is worth zero
2. A small number of real companies that graduate to the Russell 1000 fast. Russell 2000 holders won't get the upside from the true long-term home runs
3. Many small, okay-ish banks. This is the best part of the index by far
4. A bunch of boring low ROCE businesses that have done nothing for years and pay a 1.5% dividend
I already have a post planned in my mind. "The Russell 2000 is back to where it was when I first started investing in 2018." It was $170 back then.
In real terms, it's already below that level.
This Post is from an account that no longer exists. Learn more
NoEdgeMonkey retweeted
1) Talking to my family about public real estate is always a funny experiment
I have given them lots of ideas that doubled in 3 years which equates to 26% IRRs and batting average of over 90% (only giving them high conviction ones)
So probably low 7 figures of total profits in the past decade
Worst outcome is usually just dead money for a while
NoEdgeMonkey retweeted
Complete batshit lunacy on the president's Truth Social account. Yes, he really posted this.
NoEdgeMonkey retweeted
.... here we go
Rumor going around that $MSFT is paying penalties to get out of some data center contracts & power agreements-- changing tone drastically on CAPEX growth. Might be another DeepSeek shockwave to the semi supply chain.
NoEdgeMonkey retweeted
ChatGPT Plus increased my investment research productivity by 20-30%.
ChatGPT Pro, which I’m still learning to use, has quickly led to a 100%+ step function improvement in productivity.
Financial modeling and expert/qualitative research calls aside, ChatGPT Pro is already better than most hedge fund caliber analysts.
NoEdgeMonkey retweeted
All these recruiting companies are absolute garbage. Aerotek, Express Employment, etc. Pricing is stupid dumb and they can’t source any good blue collar candidates.
Today someone from Integrity Staffing called me bragging about their AI capabilities, and I lost it.
Me: “You guys truly think you can find plastic extrusion candidates on the Internet? Using AI is how you are going to find pple that can melt resin in a hopper, pull the bubble up 3 stories and adjust the speed to get the right gauge? You pple shld honestly camp outside the night shift at my competitors and ask those leaving at 3am if they would be interested in a day shift so they can spend time with their families. This AI shvt makes me absolutely confident your team has no faking clue what you are doing. The pple I’m looking for aren’t going to be found with AI!”
NoEdgeMonkey retweeted
this is an excellent thread that makes some really good points
Agent Meets SuperApp
Where do AI agents fit best - chatbots, browsers, OS, social platforms (e.g. WhatsApp), or something else? The Tencent x DeepSeek integration just highlighted another powerful option: SuperApps.
WeChat is the ultimate SuperApp - every consumer internet company dreams of replicating it. It seamlessly integrates messaging, social media, mobile payments, enterprise services, and millions of MiniPrograms into one ecosystem. With 1.3B DAU spending 100+ min/day, it functions like an OS but with built-in social + transaction capabilities, making it a lot more powerful.
Some Possible Agent Use Cases in WeChat:
🔹 Newsfeed Agent: WeChat’s public accounts are a fusion of news, blogs, X, and Substack. They’re the go-to information hub for 1B+ DAU. Agents can auto-generate summaries, track trends, and build personalized knowledge bases.
🔹 Social Agent: Imagine an AI that highlights key insights from daily conversations - summarizing group chats and extracting to-dos. What if every WhatsApp group had an AI daily digest?
🔹 Enterprise Agent: WeChat’s enterprise accounts = WhatsApp Business in LatAm/India. Enterprise AI could manage customer service, automate marketing, and personalize engagement at scale.
🔹 Transaction Agent: WeChat’s MiniProgram ecosystem is basically an internal App Store - users can order food, book rides, or shop without leaving the app. Agents could seamlessly handle transactions, recommend purchases, and optimize workflows.
🔹 AI Search: WeChat search lags Baidu 3:1 in query volume and earns only 1/70th of Baidu’s search revenue. With the Tencent x DeepSeek integration, it should unlock more sophisticated, agent-driven discovery and bring in search ads for Tencent.
Agent + SuperApp: The Power of a Closed Ecosystem
One of the biggest challenges for AI agents is accuracy - more steps mean more chances for errors. Stringing together multiple actions compounds these risks. The best setup = fewer steps, maximum standardization, and tight ecosystem integration.
Zuckerberg has big ambitions for $META AI agents - integrating them into WhatsApp Business for automated interactions, AI avatars for advertisers, and agent-driven optimizations to boost ad ROI. Given WhatsApp’s scale, these AI agents could meaningfully enhance engagement and monetization.
Tencent, meanwhile, is equally - if not better - positioned to execute on a similar roadmap. WeChat is already deeply embedded into commerce, payments, social, enterprise services, and more. AI agents don’t have to bridge fragmented ecosystems; they operate natively within WeChat’s closed-loop, high-frequency environment.
China’s walled garden model - where everything happens inside integrated, closed-loop platforms - creates fewer steps, fewer errors, and a more controlled environment. As AI agents take on more tasks, this structure becomes an undeniable advantage.
🤖Agent + SuperApp = SuperAgent.
