Loading activity…
@deedydas
Partner at @MenloVentures. Investor: Anthropic, OpenRouter, Modal, Wispr, Pangram, Inception, Goodfire, PrimeIntellect prior: Glean, Google Search. Cornell CS.
AI
The launch post claims the diffusion LLM outperforms GPT-6 Luna and Haiku 5.5 on quality.
AI
Greptile names Menlo Ventures as lead investor and says it reviews over 1% of GitHub pull requests daily.
AI
Glimpse’s team claims competitive frontier performance compared with Astra on key public visual reasoning benchmarks.
AI
Goodfire researcher Nick Cammarata argues that modern tools reveal most neurons as legible in models up to trillions of parameters.
AI
Modal says it recently crossed $1 billion in annualized run rate, with Sandboxes alone at a $300 million run rate.
AI
In October 2026, a post put the median age of the startups involved at four years.
AI
A user argues that founders can raise money more easily in the US with less dilution, while engineers earn more.
AI
A poster says the teen left a top-10 US college after two weeks, following a year-long AI internship.
AI
A user estimates OpenRouter handles about 1% of total inference spending, using Opus to convert the platform's public token stats.
AI
One post questions how much revenue OpenAI keeps after partners take a cut and argues token growth matters more.
Technology
The Financial Times, citing a person familiar with the matter, links the gap to investor comparisons with Anthropic.
AI
OpenAI says it drew on an independent mathematics advisory group’s advice about how to release the results.
AI
A user says the system hashes high-amplitude peaks from short song clips, storing timestamps and track IDs for lookup.
AI
Reflection says Beam uses 23B active parameters and plans to publish its full weights later in October 2026.
AI
Menlo describes Factory as a platform for automating software development beyond code generation, with shared context and governance.
AI
A writer who says they worked at Google argues that promotion pressure can hurt top-priority projects while lower-priority ones thrive.
AI
A user argues that frontier AI models now need benchmark wins to launch, making those numbers meaningless as a quality signal. They say price is the thing to trust instead.
AI
The creator says Opus scripted the documentary, found what they describe as Creative Commons-licensed footage on YouTube and in archives, and upscaled and colorized clips.
AI
A user points to a SQLite video they describe as a seven-minute walkthrough of what the repo does, how its code is organized and the path a query takes. They say it was generated with Opus 5.5 and Gemini 3.8 text-to-speech.
AI
One user shares a video they describe as covering the essay in under 200 seconds and thanks Opus. Another says they made a version with a one-line prompt to Opus 5.5 using its “extra” effort setting.
