- Intent
- Posts
- Carta woes, OpenAI at a fork | Intent, 019
Carta woes, OpenAI at a fork | Intent, 019
What happens when the Valley doesn't love you anymore? And are you a product, or a platform?
Intent is all about helping talent in tech become more intentional with their career by becoming more informed, more fluent, and more aware about the goings-on within tech and adjacent spaces. We welcome your feedback!
Hey — quick note: on Dec. 20th, join us for AI Fluency: Master the Chain of Thought. It’s a free virtual workshop where we’ll be walking through a reasoning technique that you can use with LLMs (like ChatGPT, Claude, or Gemini) to get 10x better prompt results. It’s also the primary innovation behind groundbreaking reasoning models like OpenAI's o1, Alibaba's QwQ, and DeepSeek's R1-Lite. Understand it, and you'll understand the next year+ of new AI breakthroughs.
This is part of a new series on AI competency for people who don't yet feel like they've tapped into the upside of our new large language model overlords. Note: this is intended for non-technical audiences or technical folk who want to be better at using and talking to AI. No coding required! RSVP here.
In tech, you’re in until you’re out
Back in the heyday of Silicon Valley's venture-fueled boom, Carta and its CEO, Henry Ward, could do no wrong. The company simplified cap table management, making it indispensable for founders swimming in equity grants and secondary market dreams. (cap table management = helping startups keep track of equity ownership; secondary market = when an existing shareholder sells to an outside investor).
Ward’s star rose as Carta became the darling of a cash-rich tech ecosystem, even spearheading an ill-fated secondary market product designed to make private equity liquid.
In fact, it was so popular that venture capitalists began to force usage of the software as part of their term sheet offers to startup founders (incl. to the startup writing what you’re reading!) — despite price tags that were astronomical compared to alternatives.
But the shine wore off. Reports of gender discrimination lawsuits and allegations of a retaliatory firing against a former CTO darkened the skies (TechCrunch). The secondary market experiment failed to deliver and caused an uproar on Twitter, and operational missteps had customers tweeting about its alternatives.
And a few days ago, a founder’s tweet on the company’s arduous cancellation process caused a stir within tech circles, with competitors like Fidelity, Pulley, and AngelList being… aggressively… recommended in the comments.
This is a classic case of Silicon Valley’s tribal ethos: one minute, you’re a hero and part of everyone’s come-up; the next, a cautionary tale. Ward, once lauded as a visionary, is now a polarizing figure. He’s seemingly no longer allowed to move fast and break things, at least in the eyes of his founder peers. In an ecosystem obsessed with rapid growth and doing things that don’t scale, Carta’s story reminds us that you can only have a tough-to-cancel product if the in-crowd still loves you.
Editor’s note: Carta is no longer forced on founders by many relevant venture capitalists. But it’s still up to the founders to figure out how to cancel a product they might not have wanted in the first place.
Platforms or Products
A few days ago, OpenAI released their video generation model, Sora — and the most interesting part to us: the clear decision they’ve made to go product-first with video generation rather than platform-first.
It comes equipped with a mini video editor, a storyboard feature, a gallery for managing generated content — it’s a product suite with no available API for third party developers, a notable difference from what OpenAI usually offers when it releases a new large language model. When a new GPT-X model is released inside ChatGPT, they either immediately or soon-thereafter release it via API.
Why the change? It could come down to trust and safety concerns. Video generation raises complex ethical questions around content authenticity, deepfakes, and misinformation. By controlling the experience end-to-end, OpenAI can set boundaries, ensure compliance, and avoid lawsuits that might have come if third party developers had gone… experimental… with things.
Beyond that, this allows OpenAI to shape the tool and pitch it as a positive augment to creatives in the media industry. They know that societal pushback is an existential threat, and being able to shape the tone and approach of a product that could “replace” filmmakers means a more considered and PR-friendly strategy.
Very few companies can be both product and platform, though. Building products means capturing the revenue directly and shaping the user experience (ChatGPT). Being a platform means rapid scale of use cases and usage (GPT-4 via API). And the needs and internal structures of companies in either lane look quite different.
And this same question likely gets re-litigated every few months and with every product release at Google, Anthropic, and Meta, too. It will be interesting to see these companies navigate the maze, especially in light of the fierce competition they find themselves in with… each other.
PS — in a job search? Join us on Dec. 19th for Behind the Curtain: How Recruiters Actually See Your Profile. We’ll be walking through a live view of an applicant tracking system (ATS) and LinkedIn Recruiter. RSVP here.
Do you know who Speed is? (aka IShowSpeed)We're curious! And we want to know. Does our tech-y audience know who this is? We'll talk more about why we're asking in the next issue of Intent. |
Alright, that’s all for now. Happy Friday!
Sent with Intent,
By Free Agency