The Frontier

Marty Bent emphasizes avoiding vendor lock-in, noting he primarily interacts with his Hermes agent via Telegram for six months, only using desktop applications for debugging.

Marty Bent details a workflow where his Hermes agent, Martin, orchestrates tasks for other agents (Fizz, Bumble, Honey) within Buzz, collaborating on company initiatives for up to 45 minutes to complete a task.

Much of the AI and surveillance infrastructure used by immigration enforcement, including Palantir technologies, was implemented during the Biden administration.

San Francisco's Grit Robotics designed AI to train Kawasaki robot arms for unloading, transporting, and initiating utility-scale solar panel installation, with humans for final fastening. This automation enables efficient "factory in the desert" solar farm construction.

Apple sued OpenAI for stealing trade secrets, alleging the theft of hardware designs and IP by former Apple employees who joined OpenAI, including an engineer who accessed Apple servers.

The lawsuit claims OpenAI encouraged new hires to study confidential Apple materials and bring hardware prototypes, prompting Apple to declare "thermonuclear war" similar to Steve Jobs' stance on Android.

OpenAI users experienced significant token burn with GPT-5.6 Soul, leading Tibo from OpenAI to temporarily remove usage limits for Plus, Business, and Pro plans and announce efficiency improvements.

Satya Nadella, Microsoft CEO, argues businesses "pay for intelligence twice" by revealing proprietary knowledge, advocating for distributed learning infrastructure to give firms control over their AI learning loops.

Kevin Roose notes that for companies like NVIDIA, open source models are beneficial as they lower intelligence costs, driving demand for chips; frontier labs also rely on open source advances for their own models.

Casey Newton suggests Substack's AI detection aims to protect its network's value, believing users won't pay for subscriptions primarily consisting of AI-generated "slop" over human writing.

Melisa Tokmak founded NETIC to build AI for large essential service businesses, including HVAC, plumbing, and pet care, acting as an autonomous layer between companies and their millions of customers.

NETIC agents autonomously handle customer interactions (calls, texts, online scheduling) by understanding complex operational needs, customer value, and technician specialties to optimize service deployment.

NETIC replaces hundreds of human support staff in essential service companies, enabling growth and improving EBITDA margins for these often private equity-owned businesses that traditionally struggle with labor reliability.

Over 70% of NETIC's customers are "AI-first," meaning their initial customer interactions with the service company are fully managed by NETIC's AI agents.

Prior to founding NETIC, Melisa Tokmak was a director of engineering at Scale AI, which later secured an approximately $30 billion agreement with Meta, where she built government and large enterprise business units.

Melisa Tokmak chose to build NETIC as a product company rather than pursue an AI roll-up strategy because her expertise lies in engineering and product development, aiming for scalable solutions applicable across many businesses.

Melisa Tokmak believes large AI labs are not competitive threats to NETIC, as labs focus on generalizable problems and tools, lacking the specialized product orchestration and "last-mile" solutions crucial for specific enterprise needs.

NETIC's North Star is to build an autonomous enterprise where its platform handles every operational aspect except the physical labor, allowing businesses to maximize focus on human service quality and customer delight.

Tokmak challenges the misconception that essential service industries are "old school," noting that many are tech-forward and value-focused, quickly adopting solutions that demonstrate clear ROI, like NETIC's.

NETIC has already generated over $600 million in net new revenue for its customers through AI-handled interactions, demonstrating a tangible return on investment beyond just cost-cutting.

Decagon shifted the majority of its AI stack to open-source models, primarily for latency optimization, enabling voice agents to deliver fast responses for large enterprises with millions of customers.

Ashwin Srinivas explains that while frontier models excel at broad, open-ended tasks like trend analysis or variant creation, Decagon uses fast, smart models for well-defined auxiliary tasks within its primary conversational flow.

Decagon's research team fine-tunes open-source models, an expensive and non-trivial process requiring custom data, benchmarks, and evaluation sets tailored to specific tasks and end-to-end customer outcomes.

Enterprises will eventually adopt fine-tuned open-source models for scaled, solidified use cases due to latency and cost benefits, but the transition is slow due to model risk governance, security, and internal inertia, according to Jesse Zhang.

Ashwin Srinivas views Decagon Labs as a 'model factory' that compresses the time between new model releases and the deployment of useful, fine-tuned models for their specific tasks, adapting to the rapidly changing AI landscape.

Decagon's agent, Duet, acts as a second, smarter AI agent designed to automate the entire process of writing agent operating procedures, creating system integrations, generating tests, and monitoring conversations, tasks previously performed manually.

Jesse Zhang and Ashwin Srinivas see AI agents becoming the 'front door' of a business, handling all customer interactions, whether reactive or proactive, enabling companies to provide personalized experiences at scale.

Decagon's forward-deployed engineers are primarily focused on product improvement, translating customer needs into core product features usable by all, rather than offering one-off consulting services.

Jesse Zhang states that customer support, though a core initial use case, revealed an agent capability for following business processes, enabling Decagon to expand into inbound sales and operational workflows as models improved in instruction following.

Ashwin Srinivas highlights that for many enterprises, increased AI efficiency in customer support (e.g., a 30% cost reduction) often leads to expanded service rather than layoffs, due to previously unmet demand.