Alex L

@ian_dot_so OAI has flex tier too 📈

I bet we’re ~6mo from a vibe shift back to OpenAI GPT5.5 is very impressive and 40% cheaper, limits are higher. Codex is stellar and taking mindshare of the top devs I know. Teams I talk to are disillusioned with ROI from soaring Anthropic costs. It’s becoming untenable.

@kenwuuuu lol

so what is the current industry standard to build agents? do people still use langchain/langgraph?

Insane how there's one (1) insight you need to be good at applied AI and 99% of people completely miss it

@Lux_Stella_ In 3 years we went from “it can’t do basic algebra” to “actually this proof only requires deep knowledge of graduate-level math and the intuition to challenge conventional wisdom, it’s not a true breakthrough”

there's some really interesting commentary from one of the mathematicians in the companion paper not to pour cold water on the results necessarily - but it suggests this is a specific "kind" of proof these models might be particularly good at

bro it isn’t generally intelligent bro its only read every book and paper ever written and just making connections between them bro. its only thinking for twenty hours bro it’s just brute force thinking bro. its only solving erdos problems bro it could never be an accountant bro

@googleaidevs @RamP Hey I think I know that guy

@holland_tom Yooooooooooo

Next stop…

Chicken is the moat. Beans are the orchestration layer. Cheese is a high-signal taste indicator. Corn is orthogonal to salsa, which will see diminishing returns as compute scales. Guac is the system of record to which all of the value will accrue.

@tszzl How do I invest in a tweet

one take that always ages really badly is that the models are intelligent enough and nobody can really notice the improvements anymore

@pmarca The phrase "New York financier" has been ruined for me

The London-New York Axis of Cope. 🫣

@hwchase17 @RampLabs @RamP Holy W

Talked to @ramplabs Head of Applied Research Alex Shevchenko on the Max Agency podcast to learn how @Ramp Sheets was built, their internal agent Inspect, and so much more. YouTube: youtube.com/watch?v=trEM9O… Apple: podcasts.apple.com/us/podcast/how… Spotify: open.spotify.com/episode/49NvGQ…

@RampLabs wow!

Introducing my new model SubLinear. It's built on O(1) super-sparse attention. The context window is 1 billion and inference costs .000001 cents per million tokens. Also it gets 95% on SWE-Bench. No you cant try it out but please join my waitlist.

@mattturck @RampLabs @tryramp thanks for having me!

.@RampLabs (the AI unit of @tryramp) has been *cooking* with agentic innovation Here's @a_levitator discussing and demo'ing code self-maintaining software and the concept of AI software factories #DataDrivenNYC ______________ 00:04 - Intro 01:11 - The shift from writing code to code maintenance 01:59 - Introducing Ramp Inspect, the background coding agent 03:05 - The first experiment: Nightly AI code automation 04:23 - The limits of stateless monitoring in large observability surfaces 05:47 - Using Datadog monitors to give the AI state and focus 07:23 - Real-world example: AI autonomously fixing an authentication bug 08:14 - How to control noise and implement an AI triage pattern 09:27 - The old vs. new paradigm for continuous code observability 10:21 - Key learnings on building autonomous AI software factories

@BigBreakfastLob thinking about adding this property to my portfolio

Corcoran is listing newly renovated Flatiron building and you are looking at ~4000 dollar per sqft which is a bit insane because this is same tier as central park tower.

@blip_tm LLMs are not messy and random on saturated benchmarks. Just need to wait for SWE to become saturated

there are logical guarantees on a compiler’s correctness. if assembly comes out wrong, it’s a bug in the compiler. LLMs are messy and random and make mistakes! they’re powerful but definitely not the same thing!

@JoshPurtell Me clocking into my job at the antipattern factory:

If you wear one of those hats with upside down letters on it I should be legally allowed to kill you

When I’m in a holding competition and my opponent is the scaling laws