most people would prefer a jury of Claudes to a jury of their peers. so would I!
but would you ACTUALLY trust Claude to be your lawyer? (I wouldn’t. Not yet!)
models passed the bar exam years ago, but surviving a Tuesday in court is a very different thing
real legal work is messy. clients are vague. case files are enormous. the right answer is never black and white.
we built DAYJOB: Law to test if agents can work like the lawyer you want.
It has:
→ 71 real-world legal assignments
→ 1 GB of messy case files
→ Underspecified, one-line requests that sound like actual clients
no multiple-choice questions! no neatly packaged hypotheticals! just the deep legal reasoning you want when your heart’s pounding and you're one sentence away from behind bars.
an example task: an agent was asked to draft a sentencing memo for a client charged with multiple shoplifting incidents.
it incorrectly stacked the sentences, overstating the client’s potential prison time.
⚖️ DAYJOB: LAW, our new benchmark measuring whether agents can handle the work of experienced lawyers ⚖️
An investor is about to put $20 million into a startup. The company’s ownership spreadsheet looks clean. But its official stock ledger shows 350,000 shares held by a former executive that the spreadsheet leaves out. The board approved a buyback, but there’s no proof it ever happened.
Would an AI agent notice? Better question: would it do anything about it?
That’s one of the scenarios in DAYJOB: Law, our new benchmark for evaluating AI agents on real-world legal work, with 2,638 messy case files and nearly 1 GB of data to sort through.
The agent has to recognize that the missing shares could change the investor’s ownership stake and the price they should pay, even though the request never mentioned them.
Congrats 🏅GPT-6.2 Astra, 🥈GPT-6.2 Sol , and 🥉Sonnet 5.5 for the top three spots.
Konpo retweeted
⚖️ DAYJOB: LAW, our new benchmark measuring whether agents can handle the work of experienced lawyers ⚖️
An investor is about to put $20 million into a startup. The company’s ownership spreadsheet looks clean. But its official stock ledger shows 350,000 shares held by a former executive that the spreadsheet leaves out. The board approved a buyback, but there’s no proof it ever happened.
Would an AI agent notice? Better question: would it do anything about it?
That’s one of the scenarios in DAYJOB: Law, our new benchmark for evaluating AI agents on real-world legal work, with 2,638 messy case files and nearly 1 GB of data to sort through.
The agent has to recognize that the missing shares could change the investor’s ownership stake and the price they should pay, even though the request never mentioned them.
Congrats 🏅GPT-6.2 Astra, 🥈GPT-6.2 Sol , and 🥉Sonnet 5.5 for the top three spots.
Konpo retweeted
we built the fastest furniture search on earth and opened it to everyone.
u just talk-it's so easy my 6 yr old tried it LOL
you'll think the video is sped up. It isn't :)
Konpo retweeted
Linearity by @konpo
Generate once, scroll forever.
→ lenis.dev/showcase/linearity
Linearity
we built HANDBOOK.md: a new benchmark for long-context agentic instruction following.
we drop an agent into a live company environment with files (PDFs, Excel, Word Docs, ...), tools (email, Slack, Jira, calendar, ...), and a dense corporate handbook (up to 124 pages), across 5 enterprise domains
the agent is given one instruction: follow the company rules.
HANDBOOK.md models the way enterprise employees have to adhere to corporate policies in their everyday work, and every frontier model fails >75% of the time.
they fire employees without authorization
they clear self-"approved" expenses
they submit expired records to insurers
...and then they report full compliance.
Blog post: surgehq.ai/blog/handbook-md
Github: github.com/surge-ai/handbook
Benchmark Leaderboard: surgehq.ai/leaderboards/hand…
Most of the time, I think about AI solving everyday problems.
AI that can do my taxes.
AI that can parse a goddamn pdf!
But I also think there's a poetry in what we’re all building. Models writing Nobel Prize-winning literature that makes us think. Models that could help us understand the nature of the primes.
I think moonshots are important too.
When I was a kid, I wanted to be a mathematician. Number theory or topology. One of my favorite problems I worked on in school was the Aanderaa–Karp–Rosenberg conjecture - a graph-theoretic problem in graph theory with surprising connections to algebraic topology.
Wouldn't it be cool if models could discover those connections too?
5 years ago, we worked with OpenAI to create GSM8K, the first math reasoning benchmark for LLMs. At the time, GPT-3 couldn't break 20%... But the frontier has moved!
So I'm excited to introduce Riemann-bench, a new benchmark for moonshot mathematics that we built in collaboration with leading mathematicians around the world - Ivy League professors, PhD IMO medalists, and graduate students at the top of their field.
Check it out! surgehq.ai/leaderboards/riem…
Konpo retweeted
Design studio @konpo has rebranded the world’s largest mental health provider. The new logo is a glowing core with expanding ripples, representing growth, support and reach.
Konpo retweeted
ComPsych Design Guidelines by @konpo
Brands look better with Lenis.
→ compsych.konpo.co/
Konpo retweeted
ComPsych supports 130 million people worldwide. @konpo’s rebrand has given the mental health giant an identity to match its scale.
Konpo retweeted
Replying to @TweetsByTBI @konpo
Loved this? Get more curated insights and free resources: t-bi.link/newsletter
Konpo retweeted
Kudanil by @konpo → maxibestof.one/websites/8015…
Konpo retweeted
New studio feature on @thelandinglove today: Konpo Studio by @konpo
Konpo is a full-stack design studio—a plug-and-play crew of seasoned designers for CXOs who understand design is an unfair advantage.
Live at konpo.studio/
Konpo retweeted
New Website in Inspo → Konpo Studio
nocodesupply.co/item/konpo-s…
Konpo retweeted
New sites added to a1.gallery this week ↓
Featuring: @elliottmangham, @konpo, @jeanmassad, @Holographikco, @StudioAgri, @CarlBeaverson, @jonasbrinkhoff, @_summation, @muzaffarlii, @usebland
Konpo retweeted
longwinded. eccentric. sick of VC-backed get rich quick schemes. the Michael Jordan of post-training data. profitable. an advocate for human intelligence and creativity. apparently, all of those things describe me now. only one of them matters.
forbes.com/sites/phoebeliu/2…
Konpo retweeted
Website of the MONTH is ComPsych Brand Hub
by Konpo @konpo
Congrats!! 👏👏👏
tinyurl.com/5xpdfy28
Thanks judges! 🙏
#webdesign #webdev #cssdaWOTM #CSSDesignAwards





















