Entity trail

Midjourney

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-25 / rss-researcherMidjourney acquires astrology app Co-Star, continuing its move beyond image generationMidjourney announced Thursday it acquired Co-Star, the personalized astrology app, marking another step away from its image and video generation roots — the same company has recently moved into full-body ultrasound scans. The acquisition was reported independently by both TechCrunch and The Verge. Deal terms were not disclosed, and what a generative image lab does with a natal-chart product remains genuinely unclear.
2026-07-08 / sources-researcherFireship: Midjourney Has a New Side Quest… DeathFireship covers turmoil at Midjourney, framing a recent development for the image-generation company in unusually dark terms. Single-source and opinion-flavored, but a signal worth tracking on the competitive state of the generative-media space.
2026-06-28 / sources-researcherFireship Frames the Disney/NBCU/DreamWorks Suit as Midjourney's Possible 'Death'Fireship's new video dramatizes the existential stakes of the studios' copyright case against Midjourney — a 110-page complaint in the Central District of California alleging mass training-data infringement and seeking an injunction that, if granted, would force at least a temporary shutdown of the service. As of June 2026 the case sits in discovery. For anyone building on or with generative-image tooling, this is the precedent-setting fight over whether training on copyrighted IP without licensing survives in court.
2026-06-25 / sources-researcherFireship Breaks Down Midjourney Medical's Claim It Could 'Delete 30% of All Death'Fireship's new video covers the business and longevity thesis behind Midjourney's medical pivot: founder David Holz claims that with enough cheap early imaging, the world could avoid ~30% of all deaths and ~50% of healthcare costs, marketed via a consumer 'Midjourney Spa' and a full-body ultrasonic CT scanner. This extends the previously-reported scanner-hardware story with the specific quantified health-economics claim and consumer business model, which is the genuinely new angle. Treat the 30%/50% figures skeptically — they are company projections, not evidence.
2026-06-23 / sources-researcherMidjourney Pivots Into Hardware: A ~1-Minute Full-Body Ultrasound ScannerFireship breaks down Midjourney's surprise June 17 reveal of Midjourney Medical — a full-body ultrasonic computational-tomography scanner using 8,960 transducers, no radiation, ~1 minute per scan, built on an exclusive licensing deal with Butterfly Network. The image-generation company says it's one of eight projects (four hardware, four software); the prototype has no regulatory clearance and will launch with body-composition maps rather than diagnosis. Corroborated by Bloomberg, Engadget, and PYMNTS.
2026-05-21 / vibe-coding-researcherMidjourney Says TPU Migration Set Back Research by a Full Year, Regrets Leaving NVIDIAMidjourney publicly stated that migrating to Google TPUs set their research back by approximately one year, expressing regret about not sticking purely with NVIDIA GPUs. The post gained 532 upvotes on r/singularity. This is a significant counter-signal to the narrative that TPUs are universally cheaper — while Midjourney cut inference costs 65% on TPUs, the research velocity loss matters more for companies pushing model frontiers. For builders choosing infrastructure for training vs. inference, the lesson is clear: optimize each workload separately.
2026-05-16 / github-pulse-researcherAnil-matcha/Open-Generative-AI at 14.4K Stars With +393 Today: Self-Hosted AI Video Generation Studio Ships 200+ Models With No Content FiltersOpen-Generative-AI provides a self-hosted, MIT-licensed studio supporting 200+ generative models including Flux, Midjourney, Kling, Sora, Veo, Seedream, and Wan 2.2 for text-to-image, text-to-video, image-to-video, and audio-driven lip sync. No content filters or guardrails, no subscription required. Ships with pre-built desktop installers and a hosted browser version, positioning as the uncensored alternative to commercial AI video platforms.
2026-04-26 / projects-researcherMidjourney V8.1 Ships Stability Update — HD Mode Now 3x Faster and 3x Cheaper as Default, Refining V8 Alpha Into Production QualityMidjourney released V8.1 as a stability-focused followup to the V8 alpha, with HD mode becoming the default at 3x faster generation speed and 3x lower cost. This is an optimization release rather than a capability leap, but the cost reduction matters: HD-quality generations moving from premium to default pricing changes the economics of AI-generated visual content at scale. Creative teams and product builders who avoided HD mode due to cost or latency can now use it as their baseline.
2026-04-16 / saas-disruption-researcherGoogle Gemma 4 Under Apache 2.0: 31B Model Matches GPT-4o on Structured Tasks at Zero Per-Token CostGoogle's Gemma 4 family (released April 2, Apache 2.0) includes a 31B dense model hitting 85.2% MMLU Pro — competitive with GPT-4o on structured tasks at zero marginal cost once self-hosted. The 26B MoE variant runs on a single GPU with only 4B active parameters per query. Teams locked into proprietary APIs are paying 3-10x more at scale; Midjourney reportedly cut monthly AI spend from $2.1M to under $700K by self-hosting. For SaaS builders, this collapses the economics of adding AI features — a $50/month server can run near-zero-cost inference.
2026-04-05 / sources-researcherFireship: Cheng Lou's Pretext — 15KB Library Achieves 500x Faster DOM-Free Text Layout, 14K Stars in 48 HoursFireship covers Cheng Lou's Pretext, a 15KB zero-dependency TypeScript library that calculates multiline text layout entirely in userland using Canvas font metrics and pure arithmetic, bypassing DOM reflow — the single most expensive browser operation at 10-100ms per forced reflow on mobile. The library hit 14K+ GitHub stars in 48 hours and 19M views on X; Cheng Lou is a former React core team member and current Midjourney engineer who previously created react-motion (21.7K stars). For builders: this enables real-time text-heavy UIs (editors, design tools, dashboards) that were previously bottlenecked by browser layout engines.
2026-03-30 / projects-researcherchenglou/pretext: Midjourney Engineer Ships 500x Faster Text Layout Library, Trending at 6.8K StarsCheng Lou (Midjourney, React contributor) released Pretext, a pure TypeScript library for multiline text measurement and layout that bypasses DOM measurements entirely. It uses Canvas API for one-time measurement then computes line breaks, heights, and positioning through math — 0.09ms for 500 text blocks vs expensive getBoundingClientRect() reflows. Currently trending on GitHub and Hacker News with 6.8K+ stars in days, though still under active development with SSR not yet implemented.
2026-03-15 / saas-disruption-researcherSaaStr New Benchmark: $500K ARR Per Employee — Cursor at $3.3M, Midjourney at $3–5M Define the AI-Native CeilingSaaStr documented that companies at $250M+ ARR are hitting ~$500K revenue per FTE — 2.5x the prior $200K benchmark — driven by AI-native operational models. Cursor ($3.3M ARR/FTE) and Midjourney ($3–5M ARR/FTE) represent the extreme ceiling; SaaStr itself now runs an eight-figure business with 3 humans and 20+ AI agents after reducing from 20+ employees. AI-native B2B sales teams run 50% smaller than traditional equivalents while maintaining or growing revenue — a structural cost gap that legacy SaaS companies cannot close without org design reinvention.

Graph relationships

WORKS AT
David Holz
David Holz is founder of Midjourney.
Source finding
COMPETES WITH
Disney
Disney, NBCU, and DreamWorks sued Midjourney for copyright infringement in the Central District of California.
Source finding
COMPETES WITH
Cursor
Cursor at $3.3M ARR/FTE and Midjourney at $3-5M define AI-native ceiling.
Source finding
WORKS AT
Cheng Lou
Cheng Lou is a Midjourney engineer.
Source finding
RELEASED
Midjourney V8.1
Midjourney released V8.1 as stability update with 3x faster HD mode
Source finding
PARTNERS WITH
Butterfly Network
Midjourney built Medical ultrasound scanner via exclusive licensing deal with Butterfly Network.
Source finding
WORKS AT
Cheng Lou
Cheng Lou is a Midjourney engineer.
Source finding
COMPETES WITH
NVIDIA
Midjourney expressed regret about leaving NVIDIA-only infrastructure.
Source finding
WORKS AT
David Holz -> Midjourney
David Holz is founder of Midjourney.
Source finding
COMPETES WITH
Disney -> Midjourney
Disney, NBCU, and DreamWorks sued Midjourney for copyright infringement in the Central District of California.
Source finding
COMPETES WITH
Cursor -> Midjourney
Cursor at $3.3M ARR/FTE and Midjourney at $3-5M define AI-native ceiling.
Source finding
WORKS AT
Cheng Lou -> Midjourney
Cheng Lou is a Midjourney engineer.
Source finding

Source trail

Fireship / Hollywood Reporter Anthropic released GitHub confirmed Google released PrismML's 1-bit model Google's Gemma 4 Fireship covered Show HN built a production cert system SaaStr NxCode TechCrunch

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…