
Managing Partner @ExceptionalCap | Former Athlete
San Francisco, CA
Joined December 2021
- Tweets77
- Following272
- Followers126
- Likes94
In the words of the great @Ludacris “I feel like your every imperfection is perfect”. The power of conversation comes from empathy where you feel most comfortable knowing those in the convo don’t need to be immaculate, quite the contrary!
Wow. I’ve never been impressed by voice models until now. Previous research has focused on emotion, laughing, interruption. But what makes this impressive is
1. Under-pronunciation of words
2. Stuttering / repeating of works
3. Use of filler words / trailing sounds
4. Imperfect / inefficient word choice
5. Interrupting at inopportune time
Turns out being more human means being way less perfect
I mentioned it before but @hud_evals @datavendor_ai and the team are doing incredible things! Congrats on the launch
We're announcing the public beta of datavendor.ai
We spent the last two years building the best solution for people buying and selling data, whether you're a researcher in a lab or a company trying to sell data.
Data is the fundamental bottleneck for AI progress, and DataVendor will be the infra that solves it.
Andrew Van Nest retweeted
We're announcing the public beta of datavendor.ai
We spent the last two years building the best solution for people buying and selling data, whether you're a researcher in a lab or a company trying to sell data.
Data is the fundamental bottleneck for AI progress, and DataVendor will be the infra that solves it.
Andrew Van Nest retweeted
super excited to finally launch this.
we've been quietly building this for the past few months. in order to solve liquidity for data, you have to solve a lot of technical challenges.
very proud of the team, and more to come.
xitter.cf/datavendor_ai/status/2…
We're announcing the public beta of datavendor.ai
We spent the last two years building the best solution for people buying and selling data, whether you're a researcher in a lab or a company trying to sell data.
Data is the fundamental bottleneck for AI progress, and DataVendor will be the infra that solves it.
Andrew Van Nest retweeted
Today we're publishing the first draft of the Agent Baseline Software Factory Reference Architecture!
An opinionated blueprint for securing increasingly autonomous software factories, built with folks across the community.
It's open for feedback and contributions!
agentbaseline.org/reference-…
Andrew Van Nest retweeted
New alpha just dropped, here’s your list of top performers recruiters & VCs!
exceptionalcap.com/exception…
Andrew Van Nest retweeted
Just a few hours after releasing it, over 100,000 impressions on X/Linkedin.
Have you checked out the new E100?
We built a marketplace for collaboration.
exceptionalcap.com/exception…
This started out as an ambitious project internally at @ExceptionalCap with @Marell @grahaminhoo and @fresco_inferno ... Now seeing our 3rd iteration, it's incredible to see the progression into a true product. We have predicted exceptional talent, who are buidling meaningful cos
The 2026 Exceptional 100 is now live! 🚀 For the third year in a row, we are spotlighting 100 standout engineering and product builders at some of the fastest-growing companies in tech. These are the people architecting high-impact internal infrastructure and shipping key features and products daily.
View the full list at: exceptional100.com
Fun facts: Exceptional has funded 3 founders from prior years, and members of past cohorts have gone on to raise over +$2B in total funding!
Beyond the list, the report covers our analysis of what the data shows regarding today's top technical talent:
→ Geography and education
→ Roles, skills, and personal attributes that stand out
→ Emerging trends shaping the talent landscape
→ How talent is moving between companies, and into founding teams
→ Talent migration for prior cohorts
Huge congratulations to everyone on this year's list! And remember, if you have a high conviction idea, reach out to us any time, we want to be your first supporters!
@TheNestVC @Marell @grahaminhoo @fresco_inferno
Andrew Van Nest retweeted
I’m excited to announce that @arceuslegal is launching with $17M in funding, led by @greycroftvc, with participation from @craft_ventures, @spc, and others.
As a founder, I always hated how helpless I felt working with law firms.
I went through four or five different firms and somehow the experience was always the same.
I’d be waiting on something important to our business with no idea when I’d hear back. I’d have to re-explain our business over and over again. And I dreaded jumping on calls because I knew every minute was costing me money.
We started Arceus because we believe every business deserves a better law firm.
One that moves faster, costs less, and puts the client first.
And we’re just getting started. ↓
Andrew Van Nest retweeted
I asked people: "where should enterprises go for post-training as a service?"
30 votes: Applied Compute (@ypatil125, @rhythmrg, @lindensli)
13 votes: Trajectory (@rronak_, @michaelelabd, @QuantumArjun)
10 votes: Prime Intellect (@vincentweisser, @johannes_hage)
7 votes: Thinking Machines (@miramurati, @johnschulman2)
5 votes: Fireworks (@lqiao, @dzhulgakov, @pawelg, @jamesr66a, Chenyu Zhao, @divchenko)
4 votes: Baseten (@tuhinone, @amiruci, @saltyph, @defpan, @mudithj, @maxkirkby, @oneill_c)
4 votes: Together AI (@vipulved, @ce_zhang, Chris Ré, @tri_dao, @percyliang)
2 votes: Arcanum (Rehan Rupawalla, @danielrupawalla, Krish Thawani, @samir_walji)
2 votes: Belvedir (@thezacharyyu)
2 votes: Trainloop (@jackson_stokes, @mlpierce22)
1 vote each:
- AfterQuery (@CarlosGeorgescu, @spencermateega)
- Bespoke Labs (@madiator, @AlexGDimakis)
- Engram (@dan_biderman, @EyubogluSabri, @realJessyLin)
- Mercor (@BrendanFoody, @adarsh_exe, @suryamidha)
- Radixark (@ying11231, @BanghuaZ)
- River AI (@ibab, @dsoboliev, Ievgen Soboliev, Yaroslav Nazarov)
- Unsloth (@danielhanchen, Michael Han)
People couldn't pick a company they work at or founded. Disclosure: I'm a small investor in AfterQuery, Applied Compute, Arcanum and Trajectory. I didn't vote.
Andrew Van Nest retweeted
Most browser evals look clean in a sandbox and fall apart on the live web. We wrote a walkthrough with @browserbase on how to write browser evals that hold up on the live web.
Sites update, graders reward faked outputs, and bot detection kill your runs before the agent even starts. If you're benchmarking browser agents, your scores are probably noisier than you think.
hud.ai/research/can-you-trus…
Andrew Van Nest retweeted
Browser evals are only valuable if they hold up on the live internet. Most evals break with site updates, bot protection, and faulty graders.
We partnered with @hud_evals to write about what makes a good browser task that frontier labs actually want.
Read the full post: browserbase.run/hud
Most browser evals look clean in a sandbox and fall apart on the live web. We wrote a walkthrough with @browserbase on how to write browser evals that hold up on the live web.
Sites update, graders reward faked outputs, and bot detection kill your runs before the agent even starts. If you're benchmarking browser agents, your scores are probably noisier than you think.
hud.ai/research/can-you-trus…
What sticks on @GeminiApp Call for Me is the mid-call loop: approved personal details can move while you watch the live transcript, so holds and bookings finish without you retyping the whole context.
Google says the AI-calling feature will first be available to Pixel 11 owners in the U.S. who pay for a Gemini subscription.
spr.ly/6013BG9CFZ
The unlock in @GoogleDeepMind's Private AI Compute update is persistent memory with device-held keys. Wipe-after-task enclaves were the privacy ceiling; a vault that survives across phone and laptop is when continuous agents stop being a trust tax.
deepmind.google/blog/advanci…
I’m into @harvey shipping Tenet and pushing firms to post-train their own open-weight models. Big Law owning the intelligence beats renting everything from a frontier lab.
techcrunch.com/2026/09/09/ha…
Wild that @AnthropicAI had ~950 Claude agents mine DNA for 21 hours and spot a CRISPR-like enzyme system nobody had flagged. Swarm hypothesis gen, humans still run the wet lab. anthropic.com/news/claude-di…
Still thinking about @SnorkelAI's $350M Series E at $3.5B led by @insightpartners and S32, with @lightspeedvp back in. Eighteenfold ARR to $375M on finished datasets and RL envs is the cleanest read that the money is chasing the data factory now. techcrunch.com/2026/09/22/sn…
Honestly the Sol drop that got me is the cost cut. Astra-grade coding at half the old mid-tier price is when teams stop piloting and start wiring this into the high-volume stack.
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe.
GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.
We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
Honestly Exceptional 100 is the answer I keep giving to this. Same intensity as those programs people name-drop, a lot less theater.
We are!
@ExceptionalCap 🚀🚀
Have you seen Exceptional 100?
exceptionalcap.com/exception…























