
Wow ServiceNow reported $167M in layoff costs in the first nine months of this year, $87M of it in Q3 thanks to "AI-driven efficiencies"

CEO @final_agent

Wow ServiceNow reported $167M in layoff costs in the first nine months of this year, $87M of it in Q3 thanks to "AI-driven efficiencies"

Interesting. Return of the data lake / CDP? ClickHouse? DuckDB / DuckLake?
Slop Join: joining data from multiple MCPs inside an agent context window. I actually see this a lot in GTM teams when they’re pulling marketing data together. We need a better approach for big data through MCP.

If we were starting a few years ago we'd probably consider HubSpot as our first CRM... today I'm not sure and it feels like you can go without one for even longer than before and GTM workflows are shifting dramatically at the frontier Crazy seeing that HubSpot saw a sharp dip in net new customers falling 51% over the last two quarters

Far more consequential and urgent than the previous generation of digital transformation
All, I repeat ALL, existing software needs to be rewritten. Because this software was written by humans it is both inefficient and insecure. The cost of inefficiency and insecurity in a pre-AI world was moderate, but in a post-AI world it will destroy your company. This new effort should not be a bottoms up exercise where engineers get to have an opinion about how they set up their repo or which coding assistant to use. This needs to be done in a new, top down system of record that focuses on auditability, drift detection and verification among other things. Every company needs to do this asap as today has shown. If you want help: 8090.ai

Dang... Apple pushing for this popup is probably going to kill the onboarding magic of local coding agents and instead help cloud coding adoption? macrumors.com/2026/11/16/mac…

Clearly SaaS business models need to evolve, but above scenarios don’t sound exciting 😔

Can’t help but feel like this is a prisoner’s dilemma? Take the short term money before this is commoditized and risk being directly responsible for teaching your own replacement. Or don’t take the short term money and your competitors might take it anyway. Or somehow your entire industry conspires to not help the data industry and buys some more time.
I’m realizing that some SaaS businesses could make tens of millions in additional revenue by building RL environments for AI labs. Take a KYB startup. You’ve spent years collecting company records across 150 countries, cleaning them, linking entities, and building APIs on top. You have millions of real-world investigations hiding in your database. You can turn them into tasks for AI agents: “Find the CEO of this company and cite the official document.” “Identify everyone who indirectly owns more than 25% of this business. Follow the ownership chain across France, Luxembourg, and Singapore.” “Compare this company’s onboarding application with the registry filings. What’s missing or inconsistent?” Give the agent tools to search the records and read the documents. Then use your structured data to check the answer: did it find the right entities, calculate the ownership correctly, and cite the right evidence? Now you have an environment where agents can practice millions of investigations and get a reward when they get them right. These aren’t just KYB tasks because the same skills matter for due diligence, investing, procurement, compliance, and a ton of other knowledge work. If you can show that training on your environment makes models better at these tasks, I think you’ll find a buyer for a lot of $$$

First real office again. Grateful my monitor didn’t break in the rain from earlier. @coffeeflux2 is going to bring his new dog !

What is up with this Lua psyop popping up in my feed

Looking for a design agency (or two) to work with, marketing site / brand

Sarah has always had some pretty interesting takes, excited to hear more from her regularly. Also the production value is great!!
“Do you think that getting married young has made you less ambitious?” Welcome to the first episode of SOTA Wives with Sarah Chieng (Head of DevX at Cerebras, @MilksandMatcha) x Vivian Midha Shen (General Partner at Y Combinator, @vivianmshen). In episode 1: finding your person before getting rich, turning down startup funding to finish MIT, what a YC partner sees in founders, and how a birthday post became “psychological violence.” 4:08 Finding your husband before the exit 8:48 Did we marry away our ambition? 18:04 The internet called Sarah's birthday “psychological violence” 20:47 Does your YC partner secretly think you’re doomed? 23:47 Building a startup with your ex-boyfriend 28:54 SF cat parties

Dang didn't realize codex usage resets disappear at 9pm PT

Can't wait for this era of babysitting to be over honestly
I’ve seen this firsthand: Vibe coding fails when you can’t tell the difference between a "best practice" on paper and real-world experience. Real experience means factoring in all the trade-offs, from speed and cost to safety, and making the pragmatic choice, which even the best models can’t do. Here are some examples: 1. Data migrations. The model proposed a fleet of parallel cloud functions and new insert endpoints, adding endpoint, invocation, logging, and telemetry costs. The database was the bottleneck anyway, so a single remote server would have been enough. An agent could run the migration script there, detect unexpected data formats, update the script, and continue migrating as it encountered edge cases. It would have shipped and run twice as fast at half the cost. 2. Database constraints. Models love putting every business rule and enum constraint in the database. That sounds robust, but it can make shipping painfully slow. Every new enum value or business rule change becomes another database migration. Application-level validation is usually enough. Otherwise, monitor for inconsistencies and alert. 3. Missing variables. Models often add fallback values when an environment variable is missing. It keeps the app running, which feels convenient, especially during local development. But if infrastructure forgets to set a required variable or uses the wrong name, the app can silently run in development mode. The safer behavior is to fail loudly at startup so the deployment will halt. Knowing how to code matters because you need to recognize when the model’s "best practice" is wrong and override it.

The future is heading here: agents configuring agents. Expertise, even of this meta kind, is becoming accessible through AI.
Introducing Scout. Tell it the KPI you care about, like funded deposits, and it builds the agents, learns from every conversation, tests each change, and keeps improving that number on its own. You set the goal. Scout gets you there.

Dynamic island is proving to be a winner

Excited to be cooking with @coffeeflux2 again