Twitxy Saved

Barr Yaron

Verified account This account is verified. Learn more

@barrnanas

stop trying to make webdings happen // partner @amplifypartners
1.1K Following 2.2K Followers 2.3K Posts

Latest posts 5

Barr Yaron Verified account @barrnanas
“More money is being raised by fewer investors. So far in 2026, funds over $1 billion have taken 72% of all venture money raised. Last year it was 34%. Ten years ago, it was 16%.”
Colossus Verified account @colossusmag 5d
Venture capital is dead (again).

More money is being raised by fewer investors. So far in 2026, funds over $1 billion have taken 72% of all venture money raised. Last year it was 34%. Ten years ago, it was 16%.

Even firms famous for staying small, like Benchmark and Union Square Ventures, have caved. Benchmark’s five partners raised $425 million in 2024. This June, they raised $2 billion.

When a fund has too much money per partner, they are forced to cope by either making bad deals or by taking on less uncertainty. The former is bad business, the latter is not venture capital.

The last time this happened was the 1980s. A decade when money per venture investor tripled, median returns fell from above 30% to 8%, and firms moved into buyouts and rollups. Even Sequoia dabbled in LBOs. By 1994, Marc Andreessen called Silicon Valley “dead in the water.”

In his latest essay, Colossus contributing editor @ganeumann explains what happened next, what’s likely to happen now, and which firms are best placed to win.

colossus.com/article/ventur…
View quoted post on X
Barr Yaron Verified account @barrnanas
the bitter lesson is coming for multi-agent systems, too!

instead of hand-designing agent roles, hierarchies, and communication protocols, what if we let agents learn how to work together?

exciting work showing that swarms can learn to self-organize through RL, developing emergent coordination and division of labor (and outperforming best-of-N) ‼️
Kunal Jha Verified account @kjha02 Nov 6
What if a swarm of AI agents could learn to work together on its own? 🐝

No leader. No roles. No turn order. Just one shared reward.

Our new paper s³: multi-agent RL lets LLM swarms self-organize, beat best-of-N, and scale past their training size!

scalingswarms.com 🧵
View quoted post on X
5 replies 0 reposts 61 likes 3.1K views View on X