Routine compression of the genome at about 30 to 1 gives about 23 million bytes, smaller than Microsoft Word, describing the brain's initial conditions.

Judge as of 2002-12-12. Parts to check: (a) the genome's raw information content (about 3 billion base pairs, around 6 billion bits / about 750–800 MB) compresses at about 30:1 with routine compression; (b) the result is about 23 million bytes; (c) that is smaller than Microsoft Word (circa 2002); (d) this information describes the brain's initial conditions. True: genome compressibility around 30:1 (within roughly 20:1 to 40:1) by standard methods is supported, the resulting size is around 23 MB (roughly 15–35 MB), Word circa 2002 was larger than that, and the genome (or the part relevant to the brain) is a reasonable bound on the brain's initial design information. Mostly true: the arithmetic and comparison hold but one figure is somewhat off (e.g. the achievable ratio is lower with routine compression, but the order of magnitude is right). Mixed: about half the parts hold, e.g. the size figures are roughly right but routine compression cannot get near 30:1, or the Word comparison fails. Mostly false: the numbers rest on a real redundancy figure, but routine compression gets far less than 30:1 (e.g. general compressors give about 4:1 on DNA), so the stated result overstates it greatly; or the framing that the genome describes the brain's initial conditions is unproven and stated as fact. False: the figures are wrong by orders of magnitude, or the evidence shows outright that they are wrong.

Set before the research began.

1. E1 Large-scale compression of genomic sequence databases with the Burrows–Wheeler transform Bioinformatics (Oxford Academic) · 1 June 2012

   “For example, GenCompress (Chen *et al.*, 2002) obtains 1.92 bits per base (henceforth bpb) compression on the *Escherichia coli* genome.”

2. E2 Large-scale compression of genomic sequence databases with the Burrows–Wheeler transform Bioinformatics (Oxford Academic) · 1 June 2012

   “each gave a compression level that was consistent across all levels of coverage and none was able to compress <2 bits per base.”

3. E3 Large-scale compression of genomic sequence databases with the Burrows–Wheeler transform Bioinformatics (Oxford Academic) · 1 June 2012

   “For a much larger genome such as human, this advantage disappears.”

4. E4 MPEG-G: The Standard for Genomic Information Representation Institut für Informationsverarbeitung, Leibniz Universität Hannover (slides by Jan Voges)

   “Human genome: ~3 billion base pairs x 2 bits per base = ~750 MB”

The evidence gives compression figures of about 2 bits per base only for E. coli [E1][E2] and a raw size of about 750 MB for the human genome [E4]. It gives no compression ratio for the human genome, where Kurzweil's point about repeated sequences applies, and it does not address the Microsoft Word comparison or the brain's initial conditions, so the criteria cannot be settled.