news

'You just hired a million bad employees': How the brief tokenmaxxing era delivered the opposite of what it promised

Fortune's Nick Lichtenberg argues the tokenmaxxing boom backfired: firms rolled out AI agents faster than they could manage them. Hebbia founder George Sivulka warns most companies effectively 'hired a million bad employees.'

Published 2026-07-24Source: Fortune
Fortune source artwork

Why it matters

Spend is outrunning output. Fortune cites one firm whose Anthropic bill leapt from $20k in December toward $1M in July — 50x in seven months — plus Amazon's reported $500M single-month loss as agents burned tokens with little to show.

Tokenmaxxing read

The bottleneck is management, not model price. Sivulka reckons only 1 in 100 employees can hand AI usable context. The remedy the piece lands on is routing: cheaper models for routine turns, frontier models saved for core work.

Source takeaway

UBS now calls token-cost anxiety a real concern for around 60% of organizations. The through-line: without sharper instructions and routing discipline, extra tokens buy more agent looping, not more productivity.

Topic links

Related projects

Tools that match this angle

#2Direct
Observability

Langfuse

langfuse/langfuse

Open-source LLM engineering platform for observability, traces, metrics, evals, prompt management, datasets, and playground workflows.

32.2K3.4KSource-available
tracesevalscosts
#5Direct
Evaluation

promptfoo

promptfoo/promptfoo

A CLI and CI workflow for testing prompts, agents, and RAG systems across models, with evals and red-team style checks.

23.8K2.1KMIT
prompt-evalscirag
#6In spirit
Evaluation

DSPy

stanfordnlp/dspy

A framework for programming and optimizing language-model pipelines rather than hand-tuning one prompt at a time.

36.5K3.1KMIT
optimizationprogrammingevals
Related feed

More source-linked context

AP News source artwork
newsAN
news

Workplaces look for cheaper AI as ‘tokenmaxxing’ fades as a corporate fad

AP reports the tokenmaxxing fad is buckling as costs climb without matching productivity. Moody's AI analytics head Vincent Gusdorf, author of a new report, says bills piled in and teams realized the tools need disciplined use.

tokenmaxxingexplainerworkplace-ai
Read note
theregister source artwork
newsT
news

Nvidia shows off Vera Rubin platform for tokenmaxxing

At an Nvidia lab briefing, Ian Buck pitched Vera Rubin as an AI-factory play: 10x the tokens per watt of a GB200 NVL72 in early CoreWeave runs on DeepSeek-R1, plus a Vera CPU said to double agent speed.

tokenmaxxingexplainerworkplace-ai
Read note
Generated Tokenmaxxing editorial thumbnail for Tokenmaxxing Didn’t Die. It Mutated
newsG
newsmedium review

Tokenmaxxing Didn’t Die. It Mutated

Gizmodo reads the Wall Street Journal's CIO Journal and finds tokenmaxxing rebranded as frontier-only mandates: Shopify bars engineers from cheaper models, while Olive founder Bill Nguyen burned 774 billion tokens in a month.

tokenmaxxingexplainerworkplace-ai
Read note