Filch

Relentless profit dodger. Pro disintermediation. Pro singularity. Following the energy/compute pfp: @rektbrands🇺🇸

[Packy McCormick @packyM]

Capitalism is just so beautiful.

$8 trillion worth of market cap coming out in support of open-weight models because it's good for their business and makes them look good...

and, totally coincidentally, it also happens to be better for us consumers and practically every business except for a couple of labs. we just get that benefit for free with MSFT and NVDA's self-interest.

it brings a tear to the eye 🥲

[Kane 謝凱堯 @kane]

This is like the Delta Force of San Francisco nimbys

[Parth Gujare @ParthGujare_]

if your GTM teams are still manually updating the CRM, building audiences, doing research, and drafting emails - you're ngmi

Ramp Revenue is our internal GTM Coworker used by >90% of all our GTM teams. background agents do real work now: managing inboxes, prospecting, CRM updates, post-call follow-ups

but automating repeatable tasks isn't the point (though our sellers do love not updating the CRM manually anymore). the real unlock is that we're encoding institutional GTM knowledge that used to live in reps' heads - playbooks, account memory, what worked for which segment, why a deal moved and making it computable

every run compounds it. every human correction teaches the system. the org's rate of learning becomes the moat, not any single agent

a few things that made it work:

→ composable context layer. give an agent a live, shared view of the account and let it write back what it learns. generic models don't know your customers

→ reliable background work needs a real harness. short episodes, memory, triggers, evals, human approval points. "ship an agent" is easy - doing real work at an acceptable quality bar is hard

→ playbooks as the orchestration layer: who qualifies, why now, the message, the channel - agents turn that into the actual work

where this goes: agents drive experimentation and orchestration - propose their own plays, kill what's not working, and get sharper with every interaction

[rahul @rahulgs]

Grok 4.5 is #1 at processing real-world invoices

At Ramp, we tested models on 150k bills submitted by actual businesses, scoring them on whether they predicted every correction a human would make

Grok achieved the highest perfect-extraction rate, beating similarly priced models Gemini Flash 3.6, GPT 5.6 Terra, and Sonnet 5.

This is a demanding long-context reasoning task. The model must infer patterns across 100K+ tokens of prior invoices, business memories, and human corrections, then apply them to new bills. The goal: zero-click accounts payable, with invoices processed correctly without human intervention.

[Ian Tracey @ian_dot_so]

This is hilarious

https://t.co/GycKCqzYD1

[Eric Glyman @eglyman]

We move billions through infrastructure secured by open-source code: Linux, the cryptography behind financial transactions, all of it hardened by decades of public review. The most trusted software isn’t the most secret.

[Etherealize @Etherealize_io]

Marc Andreessen: “This is the grand unification of AI and crypto”

“I think AI is the killer crypto app… It’s now obvious that AI agents are going to need money. It’s already happening.”

Marc explains:

“My friends, who are the most aggressive users of OpenClaw, have given their Claws bank accounts and credit cards. And not only have they done it, but it’s obvious that they needed to do it… It’s just completely obvious. The number of people who have done that today is, I don’t know, probably 5,000 or or something. But it will grow. That’s how these things start.”

Source: @latentspacepod @a16z (Apr 2026)

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb

[OSF @osf_rekt]

hoping to open @RektMkts pro to general public soon - we're working on adding more features

existing pro users - plz make sure to give us feedback and we'll try to implement asap!

if you want to be notified you can join the mailing list here: https://t.co/gZuLBAIeYN

[TBPN @tbpn]

AMP PBC Founder @AnjneyMidha advocates for a new financing program in the United States that would help AI companies compete for compute.

"One of many corporate bottlenecks is that private credit markets do not see startups as investment-grade counterparties. So when they're trying to procure compute capacity three or four years out, they just can't get those contracts."

"What we need in the United States, is a financing program that gets innovative and asks, 'How do you get the ecosystem to understand the quality of these startups? They're the engine of innovation.'"

[Hunter Horsley @HHorsley]

In 1 year, tokenized equities volume on @solana has grown from:

$1.34 million

to

$3.32 billion

Capital markets are coming onchain.

Check out the report —

[Filch @__filch]

Tf else they gonna do

[Anthony Pompliano 🌪 @APompliano]

Wow.

Google is willing to risk the entire company to win the AI race

[Meltem Demirors @Melt_Dem]

narrative violation: crypto native founders are building the compute frontier

neoclouds @CoreWeave @CrusoeAI @Hut8Corp @TeraWulfInc @togethercompute

token markets @OpenRouter @PrimeIntellect @hyperbolic_labs

agent / harness @NousResearch @AskVenice

more to come 🫶

[Andrew Curran @AndrewCurran_]

Stripe is in discussions to acquire OpenRouter. Possibly for as high as $10 billion.

[amrit @amritwt]

Of course they raised $300 million

[Meltem Demirors @Melt_Dem]

dispatch from Crucible Compute

an under discussed reason for the H100 smile curve (from @ComputeDesk)

deploying workloads on new hardware and new firmware is HARD and there is a shortage of low level kernel and compiler engineers to figure it out, nor can mid stage startups afford large compute orchestration teams to manage heterogenous infrastructure since it’s intermittent and expensive work

more efficient to run on H100s where you’ve already figured out how to make things work v burn expensive GPU time trying to refactor your workloads and solve firmware issues

as hardware heterogeneity continues to increase across chips, networking, and more, huge opportunity to build abstraction layer(s) for how workloads get packaged and run on all types of hardware

today there are lots of companies addressing one or a few abstractions (@SpectralCom for CUDA compilers, for example) but we expect this market to grow exponentially by necessity

if you’re building abstraction layers for both workload runtime, token pricing, workload timing arbitrage, or more, we’d love to meet you and learn more, potentially integrate you into our compute deployment as we learn by building our own little margin optimized token factory

[Kahlil Lalji @bykahlil]

Excited to announce that @naturalpay has raised a $30M Series A led by @ForerunnerVC to build payments infrastructure for AI agents.

To celebrate, we gave our agent $100K to give to you. Just sign up and our agent will send you between $5 and $10,000 on https://t.co/6M72Ne6kJM ¹

[Craig Weiss @craigweiss]

the great migration is happening
claude -> codex

[Ejaaz @cryptopunk7213]

bruh.

so cursor, meta and ramp have all announced ai model-router platforms in the last 48 hours.

i'll keep saying it - value is moving from the single-model layer to "pricing and efficiency" multi-models. these platforms are going to be very valuable.