ShiftMag

Luis Velasco works as a Forward Deployed Engineer at OpenAI. For him, code is becoming very cheap to produce, while building the system around the model is becoming critical.

When the volume of code outpaces the cognitive budget of the people asked to read it, fatigue happens.

Static runbooks break under pressure. Operational knowledge should live in executable tools instead of documents nobody can follow at 3 AM.

NVIDIA’s Igor Dmochowski explained that LLMs are no longer limited just by model size or benchmark scores; the cost of longer reasoning means runtime, context length, memory use, and throughput matter just as much.

Dana Lawson leads R&D at Netlify, overseeing engineering, product, and design. From where she sits, she sees the real shift clearly: one person or a tiny team can now cover far more ground.

Two-thirds of organizations run AI workloads on Kubernetes, but only 7% do it daily. Katie Gamanji says the ecosystem still needs better operational maturity and standards.

Figma’s Designer Advocate argued that AI can make prototypes look finished too early, which can hide dangerous assumptions.

We’re waiting for the next big AI leap - systems learning from their own experience rather than just human data - and at Infobip Shift, Google DeepMind’s Benoit Schillings said that’s exactly where the field is headed.

At 55,000 operations per second across a 20 TB sharded cluster, you stop theorizing about databases and start reading their source code.

Infobip Shift 2026 is officially open, and OpenAI’s Luis Velasco said the real work now is bringing together the data, process, and systems agents need to do the job well and keep improving.

As infrastructure gets more complex, "inside the network" is no longer enough. Users, devices, and AI agents need their own identity and only the access they need.

We're burning thousands of dollars on AI conversations that should've been a single slash command.

ShiftMag Insightful engineering content & community

OpenAI’s Luis Velasco: Code is very cheap to produce, but the system around it isn’t

Review fatigue is real. Here’s what my team did about it

Runbooks Are Dead; Make Operations Executable

LLMs Have a New Limit – It Costs More to Think Longer

Netlify CTO: “No longer should my developers be writing code”

Only 7% of Organizations Using Kubernetes for AI Do So Daily

AI speeds up prototyping, but teams still need to distinguish decisions from assumptions

Google DeepMind: AI’s Next Step Is Learning From Its Own Experience

MongoDB: One write, two logs

OpenAI Says the Real Work in AI Coding Is Building the System Around It

AI agents need their own identity and least-privilege access

You Should Bring Your Ops to Slack