Get live statistics and analysis of Spencer Mateega's profile on X / Twitter

Today, humanity is shackled by scarcity of expertise. When expertise becomes infinitely scalable, humans will be freed to tackle problems we can't even conceive of today. Introducing @AfterQuery. We’re building a world where expertise is abundant. Domain by domain, profession by profession, AfterQuery is crafting datasets that encode excellence into forms that machines can learn. Data is the final frontier.

Introducing UI-Bench by @afterquery. The first and only rigorous eval of vibe coding tools. > 4,000+ blinded pairwise judgments > @orchidsapp, @figma make, and @lovable_dev take the lead > @v0 and @replit ranked dead last > performance gaps = differences in LLM orchestration, prompting, design templates, and post-processing > link to our paper in the comments!

4:29 PM: Penn notifies students of President Liz Magill's resignation. 4:30 PM: @kalshi shows contracts on Magill's resignation trade at 90 cents/dollar (returns 11% on her resignation) so obviously I buy as many as I can. 4:41 PM: NYT reports resignation. Thanks @luanalopeslara

Excited for @AfterQuery’s UI-Bench to be the leading benchmark for UI/web design! @figma make 👀 cc: @amasad, @rauchg, @antonosika

The frontier begets the frontier. I highly recommend reading @jaminball's latest Clouded Judgement article which spells out the AfterQuery thesis (thread)

finally got the @afterquery team to touch grass

the industry standard agentic finance benchmark has calculation errors in its own answer key @_valsai's Finance Agent benchmark says US Steel's inventory turnover is 6.49 it's actually 6.55 the Finance Agent benchmark - uses incorrect accounting - has mathematically incorrect formulas - penalizes models for using up-to-date information - contains basic errors in its answers and grading rubrics our finance experts caught these errors in seconds. now they’re building private evals for frontier labs vals ai’s Finance Agent benchmark is littered with errors, but was cited in releases like @AnthropicAI's Claude for Financial Services the leading labs want to get better at “economically valuable tasks” but they’re currently optimizing against suboptimal signals. here’s why this matters →

we’re hosting the biggest applied AI research workshop + hackathon. $10K+ in prizes, talks, bounties, and free credits. hosted by @AfterQuery, @dexterity_ai, and @AGIHouseSF. may 23-24. if you’re a top student, researcher, or builder—DM me. agenthacks.org

Even though AfterQuery operates in a highly technical domain, Shravani took the time to deeply understand our product and customers (wasn't expected, but made a huge difference in her design decisions) Behind her tasteful design were hours of positioning, mood boarding, and copywriting I cannot recommend @shrawberryy more

Today, we’re pulling back the curtains. After collecting thousands of original, human-written coding problems, @AfterQuery created internal, contamination-free evals to test LLM code generation. No leaderboard tricks. No test-set leakage. Just raw task execution. Thread 🧵

🇺🇸 249 years ago, America declared that innovation belongs to the bold. Today, we're writing the next chapter—one dataset at a time. At @AfterQuery, we believe AI's future isn't just about algorithms. It's about the human ingenuity that teaches machines to think, reason, and serve humanity better. From our San Francisco headquarters, we're proud to partner with America's leading AI research labs to fuel the intelligence revolution. Our mission is uniquely American: Bringing together the brightest minds from talent hubs like Stanford AI Lab, Berkeley AI Research, Wall Street, and Silicon Valley—united to ensure AI advancement remains in the hands of those who value freedom and human potential. While others talk about the future, we're building it. Happy Independence Day from @AfterQuery! 🦅🎆 Here's to the innovators who refuse to accept limits. Here's to the American dream.

Today, humanity is shackled by scarcity of expertise. When expertise becomes infinitely scalable, humans will be freed to tackle problems we can't even conceive of today. Introducing @AfterQuery. We’re building a world where expertise is abundant. Domain by domain, profession by profession, AfterQuery is crafting datasets that encode excellence into forms that machines can learn. Data is the final frontier.

Introducing UI-Bench by @afterquery. The first and only rigorous eval of vibe coding tools. > 4,000+ blinded pairwise judgments > @orchidsapp, @figma make, and @lovable_dev take the lead > @v0 and @replit ranked dead last > performance gaps = differences in LLM orchestration, prompting, design templates, and post-processing > link to our paper in the comments!

4:29 PM: Penn notifies students of President Liz Magill's resignation. 4:30 PM: @kalshi shows contracts on Magill's resignation trade at 90 cents/dollar (returns 11% on her resignation) so obviously I buy as many as I can. 4:41 PM: NYT reports resignation. Thanks @luanalopeslara

finally got the @afterquery team to touch grass

the industry standard agentic finance benchmark has calculation errors in its own answer key @_valsai's Finance Agent benchmark says US Steel's inventory turnover is 6.49 it's actually 6.55 the Finance Agent benchmark - uses incorrect accounting - has mathematically incorrect formulas - penalizes models for using up-to-date information - contains basic errors in its answers and grading rubrics our finance experts caught these errors in seconds. now they’re building private evals for frontier labs vals ai’s Finance Agent benchmark is littered with errors, but was cited in releases like @AnthropicAI's Claude for Financial Services the leading labs want to get better at “economically valuable tasks” but they’re currently optimizing against suboptimal signals. here’s why this matters →

Excited for @AfterQuery’s UI-Bench to be the leading benchmark for UI/web design! @figma make 👀 cc: @amasad, @rauchg, @antonosika

The frontier begets the frontier. I highly recommend reading @jaminball's latest Clouded Judgement article which spells out the AfterQuery thesis (thread)

Even though AfterQuery operates in a highly technical domain, Shravani took the time to deeply understand our product and customers (wasn't expected, but made a huge difference in her design decisions) Behind her tasteful design were hours of positioning, mood boarding, and copywriting I cannot recommend @shrawberryy more

Today, we’re pulling back the curtains. After collecting thousands of original, human-written coding problems, @AfterQuery created internal, contamination-free evals to test LLM code generation. No leaderboard tricks. No test-set leakage. Just raw task execution. Thread 🧵