TLDR Founders 2025-11-19

Modern bots behave like real users. They execute JavaScript, rotate IPs, and solve CAPTCHAs with AI. Traditional detection tools can't keep up.WorkOS Radar stops abuse at the source. Just one API call blocks bots, brute force attacks, and free trial abuse without slowing down real users.Secure your app now with WorkOS Radar.

AI is not limited to just large language models. There are other types of models, such as classification models, audio transcription models, risk analysis models, and driving models. AI has the potential to provide billions of dollars in value to the world. This article provides an overview of the different types of AI out there.

A wave of neo-labs focused on fine-tuning, tools, and enterprise adoption has emerged around major frontier model providers, but many of these will ultimately be tuck-in acquisitions rather than enduring standalone platforms. Alphabet is actively pursuing partnerships to expand TPU penetration beyond its own workloads. There is significant early-stage founder activity in AI infrastructure under a broad theme of US infrastructure onshoring and leadership. Robotics is now extremely hot - many companies are raising at high valuations with limited or no product.

Platform PM myths include believing all investments must have immediate ROI and that metrics must be instantly measurable. Metrics often need time to show value, and sharing ownership across teams can drive better outcomes. Platform innovations, though less visible, play crucial roles in long-term success by enhancing scalability, efficiency, and competitive differentiation.

AI may be a necessity in 2025 to be competitive. The unit economics alone make it almost impossible to justify hiring humans for many functions. Tutoring an AI is much easier in many cases. However, an AI team is a quiet one - it can be lonely when group meetings are gone and there's no one to grab after work drinks with.

Forward Deployed Engineers (FDEs) are software engineers who embed directly with customers to build, deploy, and maintain complex technical systems in production. They are like a hybrid of a senior software engineer and a technical consultant. The role is exploding in growth because AI deployment is hard - companies need engineers who can ensure complex AI systems actually work in the real world. If you want to become an FDE, learn Python, get comfortable with AI/ML, build customer skills, be ready to travel, and target growth-stage startups.

Justworks is the all-in-one HR solution for managing your small business with confidence. Sign up for any Justworks plan and add International Contractors at no extra cost until 2027. That means paying all of your employees and contractors in one place, without the usual $39/month per contractor fee. Get started with Justworks

Ramp Sheets is an AI spreadsheet editor designed to 10x finance teams, operators, and founders. It plans, executes, and formats automatically and can be used to model budgets, forecast revenue, and reconcile financial statements. Ramp Sheets has all of the functionality of Excel and can also search the web, make targeted edits, answer questions about the data, and write formulas automatically. A short video demonstrating the editor is available in the thread.

If models keep getting better every nine to twelve months, what does that mean for software companies trying to stay relevant? Value is drifting toward the data layer, where Snowflake and Databricks already sit, while “application” companies increasingly blur into agent labs that train their own models and move down-stack.

IPO pricing looks random right now. Years without many listings created so much pent-up demand that almost anything gets swarmed, making it hard to read real investor interest. Layer on a bull market, where even shaky companies feel “safe enough,” and the signals get even messier. Banks still have to anchor pricing to peer multiples, not hype. Then algorithms jump on every small move and amplify it, which is why day one keeps catching people off guard.

The AI in Practice survey maps where adoption is happening, where the gaps are, and how teams are hardening and scaling AI in practice.

Founders rarely talk about the tension between chasing big goals and wanting an inner life that isn't ruled by striving.

The US stock market heavily relies on AI speculation, driving economic growth and political stability despite underlying weaknesses.

A first in-house recruiter makes sense if hiring more than one person monthly for a year, as it saves on costly agency fees.

AI can make unattractive unit economics 'work' and remove bottlenecks that would previously be inhibitors to scale.