Pitched Unfairly
TechCrunch Disrupt · Final round · 2014

Every file is too big, so we compress it from the middle out.

Lossless compression, rebuilt around one idea. The demo is live.

compress · middle-out

judges_pick.mov

losslessverified

Weissman score

3.8

Hooli Nucleus 2.89

01The problem

Files get bigger every year. The pipes don't.

Video, 3D, and whatever the cameras shoot next. Every byte gets paid for twice: once to store it, once to send it.

uploading

3d_dailies_final_v7.mov

Too big for the pipe

Time remainingovernight

Storage bill

Up again

02The insight

Everyone compresses from the ends in. We go middle-out.

one file, two approaches

Pied PiperStart in the middle and work outward. Same file, same lossless output, far fewer bytes.

Before

After

03Live demo

Hand us any file. Watch it shrink live.

A judge picked this one: 132 GB of 3D video, never seen by our code.

pied piper · compress

In

132.0GB

Out

26.4GB

Weissman score

5.2

Record

04Proof

Hooli's Nucleus tied our old score. Middle-out set the record.

W = α · rr̄ · log T̄log T

Compression ratio and speed, against a standard compressor. Higher wins.

  1. Hooli NucleusLaunched on this stage, yesterday
    2.89
  2. Pied Piper, beforeOur score coming in
    2.89
  3. Pied Piper, middle-outRewritten overnight
    3.8
  4. Middle-out, judge's file132 GB of 3D video, live
    5.2

05Why now

Hooli just proved the market. Now the score decides it.

  1. 01 The incumbent shipped

    Hooli launched Nucleus on this stage. The biggest company in the valley just told every buyer that compression matters.

  2. 02 The files got heavier

    The judges carried in 3D video. Every new format is bigger than the one before it.

  3. 03 The score is public

    One number, the Weissman score, that any buyer can check. A tie goes to the bigger brand. A record doesn't.

06Market

Video companies pay by the byte. We sell fewer bytes.

Illustrative, not a forecast

Beachhead, bottoms-up

2,000

video companies

$250K

license a year

$500M

a year

  1. StartVideo platforms and studios
  2. NextCloud storage and backup
  3. ThenEvery phone and every camera

07Team

Five people, one algorithm, and a $10M offer we turned down.

  1. Richard Hendricks CEO

    Wrote the algorithm. Left Hooli to build it.

  2. Erlich Bachman Board, incubator

    Founded Aviato. Holds 10% for housing us.

  3. Bertram Gilfoyle Systems

    Servers, network and security.

  4. Dinesh Chugtai Engineering

    Builds the platform around the engine.

  5. Jared Dunn Business

    Left Hooli to join. Wrote the business plan.

08The ask

Raising $5M to put middle-out in production within a year.

  • 60%Engineering
  • 25%Infrastructure
  • 15%Go to market
Seed

$5M

  1. Month 3Public API, middle-out on any file type
  2. Month 6First video company in production
  3. Month 12Three in production, record score held on their files