Offshore
Photo
Dimitry Nakhla | Babylon Capital®
This software sell-off is something to behold
A broad drawdown with massive multiple compression across an entire sector is rare — the way it happened, the speed + scale + magnitude of it has been incredible
If nothing else, it makes you a sharper investor. You walk away with a permanent reminder that multiple compression is always a risk — even for great businesses
That’s why valuation discipline matters:
You either need a margin of safety, or you need to be continuously monitoring whether growth is justifying the price you’re paying
A lot of investors rationalize:
“You just have to pay up for quality”
Sure — but at what cost? Quality doesn’t mean an infinite multiple
As Warren Buffett famously asserted:
“For the investor, a too-high purchase price for the stock of an excellent company can undo the effects of a subsequent decade of favorable business developments”
On the other hand, look back at some of the world’s best businesses and where they traded, relative to growth, over the last year:
$GOOG 18x growing >15%, $NVDA 25x growing >20%, $ASML 26x growing >17%, $MSFT 26x growing >15%, $MA 26x growing >15%, $FICO 36x growing >20%, $TDG 33x growing ~17%
Why pay 70x for a SaaS company growing 20%? It looks obvious now, but a year ago it took contrarian thinking coupled with conviction to see it that way
That’s why Benjamin Graham’s simple “2G rule of thumb” is so helpful
It keeps you grounded during euphoric periods when you’re tempted to make “exceptions” — the exact moments when investors end up overexposed to valuation risk
tweet
This software sell-off is something to behold
A broad drawdown with massive multiple compression across an entire sector is rare — the way it happened, the speed + scale + magnitude of it has been incredible
If nothing else, it makes you a sharper investor. You walk away with a permanent reminder that multiple compression is always a risk — even for great businesses
That’s why valuation discipline matters:
You either need a margin of safety, or you need to be continuously monitoring whether growth is justifying the price you’re paying
A lot of investors rationalize:
“You just have to pay up for quality”
Sure — but at what cost? Quality doesn’t mean an infinite multiple
As Warren Buffett famously asserted:
“For the investor, a too-high purchase price for the stock of an excellent company can undo the effects of a subsequent decade of favorable business developments”
On the other hand, look back at some of the world’s best businesses and where they traded, relative to growth, over the last year:
$GOOG 18x growing >15%, $NVDA 25x growing >20%, $ASML 26x growing >17%, $MSFT 26x growing >15%, $MA 26x growing >15%, $FICO 36x growing >20%, $TDG 33x growing ~17%
Why pay 70x for a SaaS company growing 20%? It looks obvious now, but a year ago it took contrarian thinking coupled with conviction to see it that way
That’s why Benjamin Graham’s simple “2G rule of thumb” is so helpful
It keeps you grounded during euphoric periods when you’re tempted to make “exceptions” — the exact moments when investors end up overexposed to valuation risk
Software! https://t.co/zLJmvedrEo tweet
Offshore
Photo
God of Prompt
🚨 Researchers just made LLMs talk to each other WITHOUT generating a single word.
Cache-to-Cache (C2C) lets AI models talk directly through their KV-Caches, bypassing text entirely.
8.5-10.5% accuracy boost. 2× faster. Zero token waste.
Here's the breakthrough (and why this changes everything): 👇
tweet
🚨 Researchers just made LLMs talk to each other WITHOUT generating a single word.
Cache-to-Cache (C2C) lets AI models talk directly through their KV-Caches, bypassing text entirely.
8.5-10.5% accuracy boost. 2× faster. Zero token waste.
Here's the breakthrough (and why this changes everything): 👇
tweet
Offshore
Photo
Brady Long
This is wild 🤯
Multi-agent LLMs are starting to think together.
A new paper "Emergent Coordination in Multi-Agent Language Models” just dropped, and it’s mind-blowing.
Researchers found that when you connect multiple LLMs with light feedback (and no direct communication), they can spontaneously develop roles, synergy, and goal alignment.
Using an information-theoretic framework, they measured true emergence moments where the whole system predicts the future better than any agent alone.
Here’s the wild part 👇
When each agent got a persona and a prompt to “think about what others might do,
→ the group started showing identity-linked specialization and goal-directed complementarity.
In other words:
> “LLM societies can evolve from mere aggregates to higher-order collectives just by changing their prompts.”
Read the full 🧵
tweet
This is wild 🤯
Multi-agent LLMs are starting to think together.
A new paper "Emergent Coordination in Multi-Agent Language Models” just dropped, and it’s mind-blowing.
Researchers found that when you connect multiple LLMs with light feedback (and no direct communication), they can spontaneously develop roles, synergy, and goal alignment.
Using an information-theoretic framework, they measured true emergence moments where the whole system predicts the future better than any agent alone.
Here’s the wild part 👇
When each agent got a persona and a prompt to “think about what others might do,
→ the group started showing identity-linked specialization and goal-directed complementarity.
In other words:
> “LLM societies can evolve from mere aggregates to higher-order collectives just by changing their prompts.”
Read the full 🧵
tweet
Offshore
Photo
Illiquid
$1024.hk - Higgisfield sells access to Kling AI.
tweet
$1024.hk - Higgisfield sells access to Kling AI.
Higgsfield is officially the fastest-scaling GenAI company in history, doubling from $100M to $200M in just 2 months.
$200M annual run rate in under 9 months.
We have raised a $130M in Series A at a $1.3B valuation, backed by Accel,
AI Capital Partners (Alpha Intelligence Capital's US-based fund), and Menlo Ventures.
To the creators and partners who built this with us: thank you for making Higgsfield a global movement. tweet
Offshore
Photo
Moon Dev
zoom call
todays zoom call is starting soon
you dont want to miss this: https://t.co/JbJdIbVuzB
moon dev
tweet
zoom call
todays zoom call is starting soon
you dont want to miss this: https://t.co/JbJdIbVuzB
moon dev
tweet
Offshore
Photo
God of Prompt
RT @alex_prompter: GOOGLE'S 42-PAGE AGENT WHITEPAPER... PACKED INTO ONE MEGA-PROMPT
I spent days reverse-engineering Google's entire agent framework into a 10-phase interactive prompt that builds production agents from scratch:
cognitive architectures, tool design, orchestration layers, memory systems, and deployment patterns.
It feels like having a DeepMind researcher walk you through building agents that actually work 👇
tweet
RT @alex_prompter: GOOGLE'S 42-PAGE AGENT WHITEPAPER... PACKED INTO ONE MEGA-PROMPT
I spent days reverse-engineering Google's entire agent framework into a 10-phase interactive prompt that builds production agents from scratch:
cognitive architectures, tool design, orchestration layers, memory systems, and deployment patterns.
It feels like having a DeepMind researcher walk you through building agents that actually work 👇
tweet
memenodes
Be like naiive, subscribed to Premium+ and show up daily to post memes. No crypto or brand promotion except his own work.
tweet
Be like naiive, subscribed to Premium+ and show up daily to post memes. No crypto or brand promotion except his own work.
X money just came. i’m crying. Elon Musk sent me 6 ETH this month.
Get X premium and start posting today.
The X algo is supporting small accounts rn . You’re gonna make it . tweet
X (formerly Twitter)
naiive (@naiivememe) on X
X money just came. i’m crying. Elon Musk sent me 6 ETH this month.
Get X premium and start posting today.
The X algo is supporting small accounts rn . You’re gonna make it .
Get X premium and start posting today.
The X algo is supporting small accounts rn . You’re gonna make it .
Offshore
Video
Startup Archive
Elon Musk on how to get people to believe in your startup
“If you think about a company, a company is a group of people that are organized to create a product or service. So in order to create such a thing, you have to convince others to join you in your effort.”
But, as Elon explains, this is hard in the beginning:
“In the beginning, there will be few people who believe in you or what you’re doing. But then over time — as you make progress — the evidence will build and more and more people will believe in what you’re doing.”
So, Elon argues, the first thing you should do when starting a company is create a good demo or mock-up so that people can really envision what it’s about.
“Try to get to that point as soon as possible, and then iterate to make it as real as possible, as fast as possible.”
A good example of this is the Tesla Roadster. Tesla debuted the concept vehicle in 2006 with a famous photograph of co-founder J. B. Straubel driving California Governor Arnold Schwarzenegger. Tesla was a long way from mass-producing the Roadster at any level of affordability, but the demo made their vision more real and helped Tesla secure additional funding. This funding bought the company more time and resources to deliver its first 500 vehicles between February 2008 and June 2009.
Video source: @ComputerHistory (2013)
tweet
Elon Musk on how to get people to believe in your startup
“If you think about a company, a company is a group of people that are organized to create a product or service. So in order to create such a thing, you have to convince others to join you in your effort.”
But, as Elon explains, this is hard in the beginning:
“In the beginning, there will be few people who believe in you or what you’re doing. But then over time — as you make progress — the evidence will build and more and more people will believe in what you’re doing.”
So, Elon argues, the first thing you should do when starting a company is create a good demo or mock-up so that people can really envision what it’s about.
“Try to get to that point as soon as possible, and then iterate to make it as real as possible, as fast as possible.”
A good example of this is the Tesla Roadster. Tesla debuted the concept vehicle in 2006 with a famous photograph of co-founder J. B. Straubel driving California Governor Arnold Schwarzenegger. Tesla was a long way from mass-producing the Roadster at any level of affordability, but the demo made their vision more real and helped Tesla secure additional funding. This funding bought the company more time and resources to deliver its first 500 vehicles between February 2008 and June 2009.
Video source: @ComputerHistory (2013)
tweet