"The AI ecosystem is on a knife edge right now where any perceived weakness, real or not, or any, but satisfaction of expectations leads to pretty significant drawdowns." Despite a healthy business, Broadcom's stock dropped after earnings. @ttunguz explains the volatile AI chip

You can’t put ERP data in a data warehouse But you can build an integrated solution that combines DWH and inventory management and finance reporting. Back at Blue Bottle I was tasked with FP&A + ML + Inventory Management and “just use redshift” was my original sin. In the

Today on The Information’s TITV: -ChatGPT 5.2 review from a legal lens | @nikogrupen, @Harvey -Broadcom reports big AI chip revenue jump | @ttunguz, @Theoryvc -@Wealthfront CEO David Fortunato on its IPO, valuing firm at $2.6B -Rubrik CEO @bipulsinha on its growth playbook

Today on The Information’s TITV: -ChatGPT 5.2 review from a legal lens | @nikogrupen, @Harvey -Broadcom reports big AI chip revenue jump | @ttunguz, @Theoryvc -@Wealthfront CEO David Fortunato on its IPO, valuing firm at $2.6B -Rubrik CEO @bipulsinha on its growth playbook

This month’s episode of 𝗜𝗻 𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗲 with @wileycwj sends @BEBischof back to his days thinking about cafe labor, warehouse ops, and the myriad of complexities in... *checks notes* ...physical goods. https://t.co/ktA8htxUIH

Theory’s AI team is spending a little time this afternoon buffing up on our fundamentals of NNs

https://t.co/aTkqPc3mkF

AI is rewriting the rules for GTM and sales leadership. Pricing models, org design, and compensation structures are shifting faster than most teams can keep up – and operators building in this environment need a new playbook. Join us for an in-person Office Hours session where

Hats off to @Theoryvc for back-to-back top-tier developer events. This hackathon was a delightful grind that yelled the quiet part out loud: there’s so silver bullet in building agents for data analysis. Just a thousand lead bullets and Claude code.

sooo apparently if you take 100 of sf's top engineers put them in a room and make them compete.. they all vibe code??

Over 100 AI Engineers built agents answer our 'quick questions'. At America's Next Top Modeler we: - created a fake company, Retail Universe, with a data warehouse, logs bucket, and over 90k pdfs. - wove 65 questions through those data - built automatic evals to test what data

This week on the podcast, we were joined by @ttunguz of @theoryvc. We talked about the major players in AI coding, the red flags we saw in the market this week, and if the access premium in the private markets has overwritten the previous illiquidity discount. We started with

🎥 Watch the full conversation with @jamiedavidson from @omni:

The products that win will be the ones that show their reasoning and give users a way to guide it.

As AI becomes the front door to data, clarity becomes the real performance metric.

When a model gives the wrong answer, the issue is usually a small misunderstanding between the question and the query.

They come from interpretation, not computation.

Most of the hard problems in AI aren't what you think.

@coledumansk19

When's the last time you evaluated your evaluator? The LLM tooling market is getting noisy, and differentiation is fading fast. How does it feel using eval tools inside a real workflow, under real constraints, with real debugging needs? During his internship, @coledumansk19