Shawn Shen is the 28-year-old cofounder and CEO of Memories.ai, a startup that builds AI to see and understand visual data. He got a Ph.D. at the University of Cambridge before joining Meta as a research scientist. Late last year, Shen left Meta to launch his startup, raising an $8 million seed round this summer backed by Samsung and others.

Meta has supercharged the Silicon Valley talent wars by making staggering, nine-figure offers for some AI researchers and starting a new AI unit, Meta Superintelligence Labs. That’s sparked tensions in its sprawling AI operations, with some Meta staff leaving.

Memories.ai announced Thursday that it’s offering up to $2 million compensation packages for researchers from Meta, Google, Microsoft, Anthropic, xAI, and others. It also recently hired Chi-Hao Wu, a former Meta research scientist, as its chief AI officer.

This is an as-told-to essay based on a conversation with Shen. It has been edited for length and clarity. Meta and Microsoft declined to comment. Google, OpenAI, xAI, and Anthropic didn’t respond to requests for comment.

Why I’m offering AI researchers $2 million

It’s because of the talent war that was started by Mark Zuckerberg. I used to work at Meta, and I speak with my former colleagues often about this. When I heard about their compensation packages, I was shocked — it’s really in the tens of millions range. But it shows that in this age, AI researchers who make the best models and stand at the frontier of technology are really worth this amount of money.

We’re building an AI model that can see and remember just like humans. The things that we are working on are very niche. So we are looking for people who are really, really good at the whole field of understanding video data.

We’re not worried about running out of money

We are welcoming people who want to take more equity compared to cash, which means that it won’t shrink our runway by a huge amount. The exact cash-versus-equity split will depend on the person we hire. We will treat these hires as founding members, not as employees. Anyways, equity is where you can get a hundred or even a thousand times return in the future.

We are thinking of hiring three to five people in the next 6 months, and another five to ten in the next 12 months. We plan to raise more money, too.

Spending so much on talent will help, not hurt, our fundraising

As long as we have the ability to consistently attract top AI talent, raising additional capital will not be a problem. The capital markets are eager to back companies that can do this. Just look at how much Thinking Machines Labs has raised or how much Fei-Fei Li’s startup has raised. As long as an AI company can recruit the best AI people, they can really just go through any kind of economic period.

Meta’s constant reorgs help our hiring efforts

Meta is constantly doing reorganizations. Your manager and your goals can change every few months. For some researchers, it can be really frustrating and feel like a waste of time. So yes, I think that’s a driver for people to leave Meta and join other companies, especially startups.

There’s other reasons people might leave. I think the biggest one is what Mark (Zuckerberg) has said: in an age that’s evolving so fast, the biggest risk is not taking any risks. So why not do that and potentially change the world as part of a trillion-dollar company?

We have already hired Eddy Wu, our Chief AI Officer who was my manager’s manager at Meta. He’s making a similar amount to what we’re offering the new people. He was on their generative AI team, which is now Meta Superintelligence Labs. And we are already talking to a few other people from MSL and some others from Google DeepMind.

I learned a lot of great things from Meta

I definitely learned a lot from Meta because Meta is very bottom-up. So you see a lot of innovations across different departments. Things like multimodal, visual, and super-personalized AI — everyone is so open to talking about their ideas. I met with so many talented people. I made an effort to meet three to four of them every week to talk about our hobbies and future goals.

It really shaped my future and gave me a clear road map. But in the end, the reason I left Meta is that I wanted to start a great company.

Amazon is about to enter the AI agent race in a big way, giving the tech giant another chance to make progress in the lucrative enterprise software market.

The Seattle-based company is testing new agentic, AI-powered workspace software called Quick Suite, according to internal documents viewed by Business Insider.

Quick Suite empowers “every business user to make better decisions, faster, and act on them swiftly by unifying Al agents for business insights, deep research, and automation into a single experience,” according to one of the documents, marked confidential.

Several companies have been given a private preview of the new technology, and Amazon recently sent out invitations for an internal beta test.

Quick Suite positions AWS to compete more aggressively in AI with agent-driven automation. Agents are the latest frontier in generative AI, designed to take action and independently use tools to complete tasks. Companies such as Google, Amazon, Microsoft, OpenAI, and Salesforce — along with a host of smaller startups — are racing to be the main providers of agentic tools.

Another crack at SaaS

This is also another chance for Amazon to make a dent in the huge market for enterprise software and applications, known as SaaS. While the company is a pioneer in cloud computing, it’s made less of a mark so far with software that runs on top of this infrastructure.

“With over 40% of business users expected to adopt Al-enhanced work environments soon, AWS is positioned to lead this shift by providing integrated solutions that help organizations — including our own — effectively deploy and scale Al agents in the workplace,” Amazon’s beta test invitation states.

Business Insider previously reported that Amazon was working on a unified “agentic” AI workspace, internally codenamed Q Business Suite, with “Quick” floated as a potential brand name.

“We are seeing strong growth of Amazon Q Business with customers like Remitly, Nasdaq, and Smartsheet, along with partners like Zoom and Asana, adopting it to provide employees with generative AI assistance to transform how work gets done,” an Amazon spokesperson said in a statement.

“Similarly, we are seeing customers like BMW and GoDaddy embrace Amazon QuickSight in order to make data-driven decisions quickly,” the spokesperson added. “We’re building on this strong response with even more innovation to help customers realize the benefits of agentic AI in the workplace.”

Merging existing AWS products

Quick Suite will merge some of AWS’s existing products, such as its data analysis platform QuickSight and its AI chatbot Q Business, while also adding a new product called Quick Flows, according to another of the documents.

Quick Flows provides pre-built workflows that let customers automate tasks through natural language prompts, one of the documents explained.

Quick Suite will include a “deep research agent” to generate reports from company and external data, and will also enable customers to create custom agents for “specific business functions or team needs,” which can then be shared across their organization, one of the documents said.

Beta testers and early feedback

Amazon is inviting a “select group of beta testers” for Quick Suite and has already offered a “private preview” to at least 50 companies, the documents said.

It’s unclear when Amazon plans to officially launch Quick Suite. The tentative launch date has already been delayed from mid-July to September, one of the documents said.

Quick Suite has received mixed feedback from beta users, according to one of the internal documents.

Customers praised its simpler setup and more intuitive design compared to Q Business, as well as its “compelling” deep research feature for both internal and external use. The ability to connect with external tools, such as Atlassian’s Jira, has also been well received.

At the same time, some testers reported frustrations with networking limitations in virtual cloud environments and onerous permission requirements for linking data sources. Early users include BMW, Intuit, and Koch Industries, the document said.

Quick Suite’s launch represents a turning point in AWS’s AI application strategy. Q Business had been intended as the company’s flagship offering for business users, but AWS is now making Quick Suite a priority, as BI previously reported.

Have a tip? Contact Eugene Kim via email at ekim@businessinsider.com or Signal, Telegram, or WhatsApp at 650-942-3061.

Contact Ashley Stewart via email at astewart@businessinsider.com or Signal at +1-425-344-8242.

Use a personal email address, a nonwork WiFi network, and a nonwork device; here’s our guide to sharing information securely.

Thrive has launched its Managed AI services.

The new solution provides mid-market businesses with a team of experts to help build their AI strategy and implement it at scale, including secure private generative AI services coupled with data and AI Assessment tools.

While IT leaders are highly concerned about using AI in business operations, most don’t know where to start or how to keep it secure. According to Gartner, more than 60 per cent of CIOs say AI is part of their innovation plan, yet fewer than half feel the organisation can manage its risks. 

To support these businesses, Thrive’s Managed AI services offer tailored solutions for every stage of AI adoption. Through a thoughtful, customer-first process, Thrive experts will help clients understand their data-readiness for AI, define goals, create clear AI use cases and identify the solutions and processes needed to meet those goals, as well as implement, manage and maintain AI systems going forward.

“Many businesses recognise the necessity of adopting AI to remain competitive, yet numerous mid-sized companies struggle with where to begin or lack the resources to implement this technology strategically and securely,” said Bill McLaughlin, CEO of Thrive. “At Thrive, we are committed to ensuring our customers are not left behind, understanding that AI can significantly impact their business. Developing our Managed AI services offering was a strategic decision and investment, enabling our customers to receive the support they need to lead in adoption while benefiting from the exceptional service and substantial ROI they expect from Thrive.”

Thrive’s Managed AI services support businesses across the AI adoption journey – from development to deployment. Key features include: 

• Use case identification: AI can apply to many tasks, and too often, businesses fail to see ROI due to loosely defined use cases that don’t support the business or have well-defined goals attached to them. Through careful discussions about their business and industry, Thrive will coach customers in identifying the best use cases to drive the most impact.
• Data assessment: Thrive ensures that organisations have the right data infrastructure and quality to support AI initiatives.
• Security implementation: Thrive has long been a leader in providing NextGen cybersecurity services. With its deep market intelligence and 24x7x365 monitoring, Thrive is bringing its security industry leadership to its Manage AI Services, ensuring data is protected and well-governed, and AI is protected from bad actors.
• AI policy and governance: Informed on the latest compliance requirements and based on industry best practices, Thrive’s team of experts will work with customers to determine how, when, and where AI is used within their business and ensure everyone follows the policy.
• Employee training: Thrive’s Managed AI experts will train employees on how to use the tools at their disposal, the pre-determined use cases the business has agreed to, and policies and best practices to ensure they are being as productive as possible, without putting the business at risk.

“From getting data sources organised and ready to ensuring the cloud capabilities are in place and optimised, there is so much technical prep that has to be done to ensure a business is ready to deploy AI and use it successfully,” said Michael Gray, chief technology officer at Thrive. “That’s why Thrive has gone out of its way to build a team of AI experts completely dedicated to just this practice. From data readiness to customisation to deployment, Thrive’s Managed AI services support businesses in the entire adoption process.” 

With the new Managed AI services, Thrive is expanding to build personalised, private AI for its customers, reducing the risk of employees inputting sensitive materials into public models. As AI continues to evolve and adoption among Thrive’s user base continues to grow, Thrive will also manage workflows to ensure all AI tools are working efficiently, as well as take on vertical model development to provide accurate and industry-specific outputs.