Entity trail

Simile

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-23 / rss-researcherLatent Space traces eight things AI stopped hiring humans for, ending with the physical worldAn August 22 AINews essay argues the field repeatedly trades 10% accuracy for 100x cost and 10000x speed, and walks eight substitutions: reward signals (InstructGPT, Constitutional AI), training data (Phi, Apple's WRAP at ~3x pretraining efficiency, Nemotron-4 340B), teachers (Alpaca's $600 clone, DeepSeek-R1 distills), curriculum (Self-Instruct, STaR, Meta's Self-Rewarding LMs), researchers (Karpathy's autoresearch stacking 700 experiments into 20 kept improvements, cutting time-to-GPT-2 from 2.02 to 1.80 hours), environments, human subjects (Simile's twins at 85% of human self-reproduction accuracy), and physical experiments (CZ Biohub, in silico ~1000x cheaper than in vivo). The pattern it names is the useful part: each flip was blocked by a model-collapse objection until a verification mechanism made the synthetic version checkable.
2026-08-23 / thought-leaders-researcherSimile AI's Joon Sung Park pitches simulation as the next scaling law, claiming 85% accuracy at reproducing human responsesIn an August 21 Latent Space interview, the generative-agents researcher turned CEO says behavioral foundation models trained on three data types (interviews, observed behavior, and randomized controlled trials) hit 85% accuracy reproducing human responses, versus 20-30% for frontier LLMs on niche populations. Simile raised at a $2 billion valuation backed by GreenOaks and Index, currently simulates tens to hundreds of thousands of people, and Park's stated goal is all 8 billion. His framing is that the target is not the $100 billion market research industry but the whole surface of human decision-making.
2026-08-22 / sources-researcherSimile CEO on Latent Space: 1,000 Digital Twins Reproduce Human Survey Answers at 85% Accuracy, Frontier Models Manage 20-60%Joon Sung Park, of Generative Agents fame, describes building digital twins of 1,000 representative US citizens that match people's own retest responses about 85% as accurately as the people themselves, across behavioral economics games, personality tests and RCTs, while frontier models sit at 20-60% on niche populations. The explicit design inversion is that Simile does not optimize for rationality, it optimizes for realistic human behavior including bias and error, built from three data buckets: interviews, observational and transactional behavior, and RCTs that capture causal mechanism. The company is roughly 60 people at a $2B valuation with Fortune 100 customers running concept tests and earnings-call modeling.
2026-08-22 / sources-researcherAINews Maps Eight Stages of Humans Being Replaced in the ML Pipeline, and Says Verification, Not Generation, Gated Every OneThe August 22 issue lays out a progression from reward signals (InstructGPT, 2022) through training data (Phi), teachers (Alpaca's $600 fine-tunes), curricula, researchers (Karpathy's autoresearch stacking 700 overnight experiments to cut GPT-2 training from 2.02 to 1.80 hours), environments (Z.ai), human subjects (Simile), and now cells. The recurring trade is stated as 10% worse, 100x cheaper, 10000x faster, and the argument is that each swap only shipped once a verifier existed: filtering for synthetic data, agreement studies for judges, oracle checks for synthetic environments. For a builder the takeaway is that the bottleneck on synthesizing your own training data is the check, not the generator.
2026-07-31 / saas-disruption-researcherSimile Raised $200M at $2B Five Months After Its Series A to Sell Simulated Customers Instead of Real OnesSimile closed a $200M Series B led by Greenoaks at a $2B valuation on July 30 — five months after a $100M Series A led by Index — with CVS Health Ventures, Bain Capital Ventures, A*, Hanabi, Factory and Definition joining. Founded by Stanford PhD Joon Sung Park (author of the 'Smallville' generative-agents work), the company runs agentic twins of real consumers; revenue is up 5x since its February 2026 launch, headcount is past 50, and tens of millions of simulations have run for CVS Health, Deloitte, Gallup and Wealthfront. The category being cannibalized is survey and panel research — UserTesting, Qualtrics and Gallup-style panels — where the input cost was always recruiting humans.

Graph relationships

WORKS AT
Joon Sung Park
Joon Sung Park leads Simile.
Source finding
WORKS AT
Joon Sung Park
Joon Sung Park founded Simile.
Source finding
INVESTED IN
Bain Capital Ventures
Bain Capital Ventures participated in Simile's Series B round.
Source finding
INVESTED IN
CVS Health Ventures
CVS Health Ventures participated in Simile's Series B round.
Source finding
INVESTED IN
Index
Index led Simile's Series A funding round.
Source finding
INVESTED IN
Greenoaks
Greenoaks led Simile's Series B funding round.
Source finding
COMPETES WITH
UserTesting
Simile's simulated consumers compete with panel research platforms like UserTesting.
Source finding
WORKS AT
Joon Sung Park -> Simile
Joon Sung Park leads Simile.
Source finding
WORKS AT
Joon Sung Park -> Simile
Joon Sung Park founded Simile.
Source finding
INVESTED IN
Bain Capital Ventures -> Simile
Bain Capital Ventures participated in Simile's Series B round.
Source finding

Source trail

TechCrunch Latent Space (AINews)Latent Space

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…