fish/family/freedom. building atas vc. fmr @BuildersVC, @nasajpl, and @caltech

Los Angeles, CA
Joined March 2013
Incredibly proud of Alex and the @GPTZeroAI team. My first exit 🥹
I'm proud that announce that today, GPTZero announced plans to join Superhuman. When we founded GPTZero in 2023, the goal was to preserve what's human on the internet. Today, that mission takes a big leap forward. Together, we're building an authenticity layer that travels with you, wherever you read, write, and create. I personally have been super impressed with the team at Superhuman and I'm very excited to join forces to start this next chapter. We are going to double down on GPTZero and its GTM, bringing our AI detector and hallucination detector to Superhuman's massive userbase, as well as creating products that help students and educators navigate what learning looks like as AI evolves. Up and onwards!
3
1
10
4,438
not that I have any reason to know this but this sounds like they used John the Ripper/similar. AI definitely makes traditional cybersecurity attacks more efficient but good security still stops anything actually important
TDLR: they got into a few Slack channels 🥱
59
Don’t know what happened to Codex usage limits but something definitely isn’t quite right right now
1
161
further in this but AI costs are only part of the equation - dollar cap makes far less sense than a token cap and model based allotment
If your employer caps your AI usage at $30 a month You do not work at a serious company
122
Saw an awful take on TikTok today that @soylent failed because their marketing reminded people of cannibalism. As an unapologetic Soylent fan and believer, I will tell you this: we used to stock Soylent at the caltech store and everyone would try it once and never drink it again because of taste. I could get past it. 99% of people can’t. Fix that and the world desperately does want a true meal replacement beverage.
1
243
Andrew Chan retweeted
I didn't even read your post, anyone who hires Quinn Emanuel to sue founders is making a deal with Satan and should be roundly shunned. xitter.cf/PalmerLuckey/status/21…
Replying to @isaiah_p_taylor
Quinn Emanuel suuuuucks. They dragged me through ten years of litigation. Good luck, brother. Fun fact, they tried to file for a protective order because they "feared for their safety" after I flipped them off (both fingers) in a Hawaiian diner. xitter.cf/PalmerLuckey/status/15…
21
14
3
760
136,529
If your own portco that you allegedly saved multiple times hates you so much they exclude you from a secondary sale and actively tries to decrease your ownership, you should probably be looking inward, not to lawyers. I’ve seen some nasty VC drama and a few even nastier lawsuits (somehow) in my career, and the truth is usually somewhere in the middle. The great news is, State of California lets you buy civil court docs so we’ll know soon enough 🍿🍿🍿 With that: unsurprising response, even if they were in the wrong at this point game theory dictates you have to double down. Realistically my bet would be on the company having done something shady (getting terms amended to screw over one investor is exactly that) but not technically illegal. Think Saverin and Facebook. Maybe a case around info rights but similarly Valar can probably claim national security for withholding that. Doesn’t change though that venture is a long game, and that even if you had every legal claim here, this is still likely short-sighted. The ROI on you suing your own portfolio company is always negative especially if they’re well-liked and have a big enough following to weaponize that against you. Especially if in the process they talk about how much they hate you.
Replying to @isaiah_p_taylor
Isaiah, after everything we’ve been through together I’m surprised that you would publicize our disagreement in this fashion. But since you have created public interest in this, it is important that people should hear both sides of the story. Day One has been a major investor in Valar from the beginning, and we have never wavered in our belief that the company will be a success. This faith and support in our founders is why fourteen Day One founders have become limited partners in Day One’s most recent fund. In 2024, Day One invested in Valar’s seed round. Later that year, Valar required a seed extension after miscalculations by management regarding the capital needs of the business. Day One stepped in to provide this seed extension, saving the company in a difficult time. When it came time to raise a Series A, Valar spent several weeks on the market without success. At Valar’s request, Day One then offered the first term sheet and co-led the Series A, again saving the company during a time when deep tech startups in highly regulated industries were a difficult sell in venture capital. Then when Valar needed a Series A extension, Isaiah asked that Day One waive our pro rata rights so as to make a tenuous financing deal work out. We agreed for the good of the company. As part of these transactions, Valar agreed that, if their long-shot effort to build nuclear energy as a startup did succeed, Day One would have rights to participate in future rounds that could not be eliminated without Day One’s consent. This was a fair bargain at a time when Day One took extraordinary risks by betting against the market when Valar’s future appeared highly uncertain. Following the achievement of criticality on June 18th, our faith has been validated. Valar went from being an obscure company to a hot startup. We congratulate the Valar team on their hard won success! Of course, when it came time for Valar to raise a Series B after achieving such success, Day One expected to participate in that round – as both parties had agreed. But at this point the normally positive communications between Day One and Valar took a negative turn. Valar insisted that Day One give up all of its future pro rata rights despite the parties’ agreements, and despite four out of six major investors participating in the Series B round without waiving. This went against both the letter and the spirit of our deal, so we did not agree. In decades of investing we have waived pro-rata rights dozens of times, but no other company has ever demanded that we give up all of our rights for all future rounds. We told Isaiah we would give up a majority of our pro-rata in the Series B round so as to not make his life difficult, to which he responded “You cannot make my life difficult, I can make your life difficult.” It was at this point that we sought representation with Quinn Emanuel to protect the firm and the interests of our limited partners. Day One has since learned of a significant secondary equity sale by Isaiah from which we were excluded. Our attempts to gain insights into this through our information rights have been stymied by Valar management, who have refused to produce material documents requested by Day One. After weeks of delay on these documents, Valar management alerted us this Wednesday that they intend to pursue legal action against Day One unless we waive our legal rights forever. We have always pursued positive and mutually beneficial relationships with all of our portfolio companies. We have neither taken nor threatened legal action against Valar, and in fact we told them explicitly that we have no interest in doing so. Valar voluntarily entered into a contract with Day One, then violated that contract, and is now suing Day One. We are deeply disappointed by this recent erratic turn, but are still proud of our early and continued support of Valar. We look forward to the clarity from the courts that Valar’s management is seeking.
7
2
30
23,281
Totally ok for Day One to ask for pro rata to be honored, totally ok for Valar to get a majority and override that. Not ok to sue because you lost a vote, ever.
18
2
1
195
32,426
Stoked for Ari and what the team at @teamairbase are cooking up… this is just the beginning
Startups are hard. You’re the youngest team in the room, the company nobody’s heard of, and there’s always more hurdles before you can help anyone. Last week the @SecWar pointed to @DIU_x Project Spectrum Strike in a memo as part of the Department’s solution to C-UAS spectrum approvals. @teamairbase won that DIU prize challenge in September. Saying that “I’m proud of my team” doesn’t even begin to describe the way I’m feeling right now. I've watched this team go through a lot to get here. The level of user-obsession is unmatched. They build for regulators, for soldiers, and for commercial operators, and they keep delivering. So yeah, I’m grateful to be able to wake up and do what I do. Now, back to work.
2
311
With @openai launching a Jev competitor 2 weeks after that launch it seems increasingly apparent to me we’re fast approaching a C2C era of software. As it becomes increasingly easy to build and ship, taste, branding, and UX/UI become all the more important
1
269
As someone who spends a ton of time web scraping this is a massive problem, and also one that I sincerely doubt companies will be able to solve without making their products a pain to use. If you want to be on Google, you’re capable. If you want to have power users, you’re automatable. LinkedIn’s solution of just rate limiting everything and everyone realistically is the only way to truly catch bot use
Bot detection & human verification will be one of the most urgent demands for businesses over the coming years. Agent swarms will suffocate every website and form; small companies and government websites are most vulnerable. There is a huge gap in the market for this right now. When we looked at what offerings were in the market to use at X, there was not a single company that brought together all the latest technologies so we had to do it all in-house.
1
1
181
The $10b valuation for Instinct imo is clearly about data. Super rich dataset, perfect for training models. Still hard for me to believe that tracks given how small a user set it is. Facebook had 5 mil DAUs when they raised A, 100k “users” is hardly the same (I get messages from Instinct daily but don’t use it daily at all)
4
176
BREAKING: VCs disclose they have been pacing their exits, declare no DPI until 2030s
15
12
131
15,847
Andrew Chan retweeted
An alternative hypothesis: - model performance is plateauing - compute is getting much more expensive - AI data centers are massively unpopular - slowing down AI development is a way to explain slowing progress, reduce spending, and try to regain some goodwill. - This is an effort to save the IPO not humanity.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: darioamodei.com/post/we-must…
271
2,342
212
20,046
915,475
a little too convenient to need to “slow down” AI when the models are still improving linearly with exponential spend increase
111
meanwhile on TikTok, the AI fear rages strong
89
anyone else remember this? let’s just not this time.
1
87
Back now - but know a few people including @_dmchan who burned a reset because of this
My codex usage just went from 50% to 10% in 10 minutes using astra medium. Think there may be some type of tracking issue? @thsottiaux
138
My codex usage just went from 50% to 10% in 10 minutes using astra medium. Think there may be some type of tracking issue? @thsottiaux
3
1
1
320
The way people talk about Instinct today feels exactly like the way people talked about Clubhouse in 2021
25
6
394
23,091