Timothy Young

CEO @heyJasperAI, Fmr President @Dropbox, Founder of Socialcast (acq by VMware), VP Prod&Eng @VMware, Founded (acq by AOL) & built lots for @Interscope

[Timothy Young @timyoung]

@levelsio @signulll We haven't gone anywhere. A lot has happened since 2022. We just closed our best quarter ever. To catch you up...

- I moved the business to the enterprise. Ended last year 163% NRR with 25% of the Fortune 500 as enterprise customers. Enterprises have a content scale problem and we help them build sophisticated AI driven content pipelines.

- We built differentiation around UX, Workflows, and Context. How you fuse enterprise data, logic, and process it consistently with an LLM is the value. Product has massively changed with a multiplayer canvas UX, our IQ context layer & integrations, and marketing agents.

- Rebuilt leadership team with execs from Stripe, SFDC, Zendesk, Amplitude, X, and Dropbox

- Opened a research lab in Paris. Acquired Clipdrop from Stability's ashes. We have some of the best image manipulation models out there and in production with companies like Wayfair, Figma, and many more. https://t.co/sF642IWJg4

[Timothy Young @timyoung]

School trained us for twenty years to find the one right answer. AI now does that part. The two skills that are left are the ones nobody taught: how to cross-examine an answer until it breaks, and how to hold a problem with ten variables without collapsing it to one.

1/ I think the two most important skills in the age of AI are Socratic Dialogue and Multivariate Thinking. Neither is taught anywhere. Both used to be optional. Now they are the job.

2/ Socratic dialogue first. Most people use a model like a vending machine. Question in, answer out, done. The people getting 10x out of these tools treat the first answer as an opening statement, not a verdict. What are you assuming? What evidence would change this? Under what conditions does it fail? What am I overlooking? That is not prompt engineering. It is the oldest intellectual method we have, pointed at a new interlocutor.

3/ That last question flips the direction, and it is the one almost nobody asks. The most valuable thing a model can do is play Socrates to you. Hand it your plan and tell it to find the contradiction before your board does. Most people never get there, because they are using AI to confirm rather than to interrogate.

4/ The second skill is harder, because school spent twenty years training it out of us. Find the one right answer. Fill in the bubble. Pick the metric. Good or bad, 1 or 0. But almost no real problem is univariate. Pricing, hiring, a product bet, an org design: each is a dozen variables that trade against each other and, worse, change each other. Ship faster and reliability pays for it. Fix one team’s incentives and the team next to it starts behaving differently. The best answer is a frontier, not a point.

5/ Here is what AI changes. A human can hold four or five variables in their head at once, so we collapsed problems to one. Not because the problems were simple, but because we were. A model can hold all twelve, map how they trade, and show you the frontier in seconds. The constraint that forced univariate thinking is gone.

6/ What it cannot do is tell you where on that frontier to stand. That is a values question, and no amount of compute answers it. AI will compute the tradeoff. It will not tell you what you are willing to trade. Which is why these two skills are really one: keep asking until the binary dissolves, then own the choice that is left.

7/ None of this is an argument for more prompting. Execution should run in systems where the model finds the work and you are not in the loop. But those systems get designed in dialogue, and what they surface is never a 1 or a 0. The skills move up a level. They do not go away.

[Timothy Young @timyoung]

@signulll The kids call this “Nowhere-maxxing”

[Timothy Young @timyoung]

I operate this way, from my first company to now at Jasper. As I tell my staff routinely: “I’m pleased, but not satisfied.”

I picked this up from David Sokol’s book of the same name. One of the best little-known business books out there (including an amazing foreword from Warren Buffet).

It teaches the founder mindset better than anything else I’ve come across.

[Timothy Young @timyoung]

The 12-month puzzle has a boring answer.

At Dropbox it was clear that utilities cross over when six things are true:

1/ The user is never a sysadmin. No setup project, no connectors, no permissions matrix.

2/ It fits mental models people already have. Dropbox was just a folder.

3/ It just works. Constant supervision is a dealbreaker.

4/ It does one thing extremely well. One repeatable magic moment.

5/ The value is trusted, predictable, repeatable. Every time.

6/ It has a sustainable emotional hook. Your folder structure was who you were and what mattered to you.

Agents today fail all six. And 5 is the killer. An agent that's right 80% of the time is not 80% as useful. It's a demo, because you still have to check the work. Predictability is the product.

When someone nails that, your friends in the group chat won't call it an agent. They'll just use it.

[Timothy Young @timyoung]

@stevesi Run into this a few times. Truly the worst. Some brands now require a dealer visit just to reset. And many won't even let you reach the battery. Land Rover's genius move: putting it under the passenger seat.

[Timothy Young @timyoung]

I spent four years as President of @Dropbox. Wall Street still prices the FCF story we built.

That story is true. It is also beside the point.

@OpenAI should buy $DBX
@AnthropicAI should buy $BOX

The file was never dead. It was waiting for a reader. ↓ https://t.co/SZVFruN3SE

[Timothy Young @timyoung]

https://t.co/8jdNjdauJ6

[Timothy Young @timyoung]

Every marketing leader I talk to is facing the same challenge: Scale 📶

The demand for content has exploded, but the systems behind it haven’t kept up. Teams are expected to launch more campaigns, in more markets, with more data, all while maintaining brand integrity and speed.

Chat-based AI tools help individuals move faster. But they don’t solve the harder problem: scaling content creation across markets and audiences.

Over the past year, we’ve seen the limits of chat-based AI inside large organizations.

📢 Today, we’re introducing ⊞ Jasper Grid ⊞ — the backbone of Content Pipelines.

⊞ Grid is a unified, no-code system that lets AI agents plan, produce, and optimize content at scale. It transforms structured inputs like briefs, data, and brand guidelines into launch-ready assets with speed and precision.

Under the hood, Grid runs on a network of modular AI agents, a resilient scheduler, and deep integrations with the tools marketers already use — all governed by brand guardrails so teams can move fast without losing control.

🏆 The result: faster time to market, consistent brand execution, and true operational scale.

AI is becoming infrastructure. The next generation of enterprise marketing will run on systems built for speed, precision, and trust.

Proud of the @heyjasperai team and our early customers who helped shape this moment.

[Timothy Young @timyoung]

Transformation is always incredibly difficult and we still have a lot more we want to do. Grinding away quietly in the shadows is something I don't mind.

I was fortunate to have learned a great playbook from Paul Martiz while I was at VMware.

But the real credit goes to the @heyjasperai team, they have done all the heavy lifting.

[Timothy Young @timyoung]

@businessbarista AI

[Timothy Young @timyoung]

I witnessed this first hand at Dropbox when I arrived.

With scale came atomization, atomization drove incrementalism, which led to distance between any one employee and the customer.

The product surface area became atomized across teams, leading people to own features instead of customer outcomes. Your job shifted to managing a feature, not fostering customer success.

- Product teams became feature-focused, rather than customer outcome-driven.

- Sales and Customer Success started gatekeeping relationships to manage retention and expansion.

- Design Research abstracted customers into personas rather than deeply understanding their evolving needs at an individual level.

- Data Science emphasized usage metrics typically driving a local maxima instead of holistic customer outcomes.

- Engineering prioritized optimization of existing behaviors to scale up.

- Legal and Security implemented rigid frameworks for customer communication and access.

- No one ended up feeling responsible for customers. The coordination costs were immense. The average employee was so booked daily internally, they had little time to even think about how to connect with a customer.

[Timothy Young @timyoung]

@wesnichols this is exactly what happened in the 2018 Woolsey Malibu fire that I lived through. Edison cut the power, which caused the pumps on PCH to stop working. Thus there was no pressure at the hydrants and no way to pump adequate water from PCH up into the neighborhoods. Local gov't failed then and obviously didn't learn anything. So sad and unnecessary.

[Timothy Young @timyoung]

🚀Quick Jasper AI update…

Over the last year, there has been a surge in demand for @heyjasperai as an increasing number of enterprise marketing teams have prioritized AI adoption. This has been hugely transformational – both for our customers' businesses, and for our own business.

Due to efforts from our incredible team, I am proud to say that over the last year, we have:

📈 More than doubled our enterprise revenue

🚀 Expanded to 850+ enterprise customers, including 20% of the Fortune 500

💰 Exceeded quota for 10 quarters straight

Our AI platform continues to lead the way with innovative solutions built for marketing success – and today, we are announcing two more product innovations. Our new ✨AI App Library, along with our powerful ✨Marketing Workflow Automation capabilities, will empower marketing teams to adopt AI faster, embedding it into every marketing process to drive impact and ROI.

Read more about our hypergrowth here:

https://t.co/z81rZNlpeS

[Timothy Young @timyoung]

VC: “Is this a B2C or B2B play?”

Entrepreneur: “…why not both?”

Most software companies early on feel pressured to choose between B2C or B2B. While focus is key, this choice can also limit long-term growth and defensibility against larger competitors.

The most powerful companies do both—when the product is valuable in both personal and professional contexts. Build an elegant B2C tool, then scale it into B2B. This opens up unique distribution opportunities, network effects, and increases switching costs.

Take @Dropbox for example, where the vast majority of users use the product for both personal and work. This is the main reason the company has been able to survive and remain competitive amongst Apple’s iCloud, MS OneDrive, and Google Drive.

Others examples? Look at @NotionHQ and @canva.

[Timothy Young @timyoung]

In practice, comparing distributed to in-office isn't straightforward. You have to redesign the company from the ground up—every process, leadership moment, procedure, and experience must be reimagined.

Having led the shift to distributed at @Dropbox (~3,000 employees) starting in April 2020, I’ve seen firsthand that it canbe more cost-effective, but it depends on many variables. For us, after deep analysis of future costs and talent considerations, going distributed was the right move. This decision was particularly challenging because Dropbox had a legendary in-office culture across its global locations.

Key cost factors to consider:

- Size of your current employee base and headcount growth projections

- Existing office footprint and its cultural significance

- Changes to your talent strategy, especially around geography and costs

- Adjustments in management and leadership approaches

- Simplification of company processes

While I agree that quarterly team meetups are ideal, the reality post-shift to remote work is that many employees found quarterly gatherings excessive. As new hires joined, they often built their lives around being fully remote.

Organizational challenges included:

- Determining the right meeting cadence (e.g., should product teams meet at quarter-end?)

- Deciding on the optimal team meeting size (pod/function/org)

- Communicating which teams were meeting and when

- Balancing the planning involved in these gatherings

Transitioning to distributed work requires intentional design thinking and planning—it's not just about choosing between an office or no office.

[Timothy Young @timyoung]

🚨 Big @heyjasperai news today….

I’m thrilled to announce that Loreal Lynch has joined Jasper as our Chief Marketing Officer! 🎉

Our executive team continues to grow and Loreal’s extensive experience in enterprise marketing and her proven track record with industry leaders like Stripe, Salesforce, and Tableau make her a perfect fit for Jasper. I couldn’t be more excited to partner with her as we continue to expand Jasper’s presence in the enterprise market.

Just a few short years ago, Jasper sparked marketers’ imaginations with the possibilities of generative AI. Today, we’re proud to support over 100,000 customers, including Home Depot, Google, Adidas, VMware, AstraZeneca, Wayfair, Harper Collins, and Cushman Wakefield, in transforming and scaling their marketing workflows.

At Jasper, we’ve learned that success with AI is a team sport. Loreal’s depth of marketing expertise and her experience in bringing category-defining technologies to market will be instrumental as we double down on scaling AI across enterprise marketing teams.

Welcome to the team, Loreal! 🚀

https://t.co/UGGipI5wjw

[Timothy Young @timyoung]

I'm thrilled to announce the launch of ⚡️Flash Diffusion from @heyjasperai!

Earlier this year, with our acquisition of Clipdrop, we launched the Jasper AI Research Lab in Paris.

Today, we are excited to release our first piece of groundbreaking research: the open-source distillation method, "Flash Diffusion".

Flash Diffusion accelerates inference by 500%, reduces computing costs, and produces higher-quality image outputs.

Dive into the details and discover how Flash Diffusion is set to revolutionize the field of AI and image synthesis.

Read all about it here: https://t.co/uZT7GUA2F1

Try a demo on @huggingface: https://t.co/j5E6CRMddF