diff --git a/go/chunks/chunk.go b/go/chunks/chunk.go
index 07e859305e..c4769114ca 100644
--- a/go/chunks/chunk.go
+++ b/go/chunks/chunk.go
@@ -34,7 +34,7 @@ func (c Chunk) IsEmpty() bool {
 
 // NewChunk creates a new Chunk backed by data. This means that the returned Chunk has ownership of this slice of memory.
 func NewChunk(data []byte) Chunk {
-	r := hash.FromData(data)
+	r := hash.Of(data)
 	return Chunk{r, data}
 }
 
diff --git a/go/chunks/chunk_serializer.go b/go/chunks/chunk_serializer.go
index 8d21a3c557..56aab704d4 100644
--- a/go/chunks/chunk_serializer.go
+++ b/go/chunks/chunk_serializer.go
@@ -31,8 +31,8 @@ import (
 func Serialize(chunk Chunk, writer io.Writer) {
 	d.PanicIfFalse(chunk.data != nil)
 
-	digest := chunk.Hash().Digest()
-	n, err := io.Copy(writer, bytes.NewReader(digest[:]))
+	h := chunk.Hash()
+	n, err := io.Copy(writer, bytes.NewReader(h[:]))
 	d.Chk.NoError(err)
 	d.PanicIfFalse(int64(hash.ByteLen) == n)
 
@@ -90,14 +90,13 @@ func DeserializeToChan(reader io.Reader, chunkChan chan<- interface{}) {
 }
 
 func deserializeChunk(reader io.Reader) (Chunk, bool) {
-	digest := hash.Digest{}
-	n, err := io.ReadFull(reader, digest[:])
+	h := hash.Hash{}
+	n, err := io.ReadFull(reader, h[:])
 	if err == io.EOF {
 		return EmptyChunk, false
 	}
 	d.Chk.NoError(err)
 	d.PanicIfFalse(int(hash.ByteLen) == n)
-	h := hash.New(digest)
 
 	chunkSize := uint32(0)
 	err = binary.Read(reader, binary.BigEndian, &chunkSize)
diff --git a/go/chunks/dynamo_store.go b/go/chunks/dynamo_store.go
index 82bc74f161..b56a17748b 100644
--- a/go/chunks/dynamo_store.go
+++ b/go/chunks/dynamo_store.go
@@ -238,7 +238,7 @@ func (s *DynamoStore) processResponses(responses []map[string]*dynamodb.Attribut
 	for _, item := range responses {
 		p := item[refAttr]
 		d.PanicIfFalse(p != nil)
-		r := hash.FromSlice(s.removeNamespace(p.B))
+		r := hash.New(s.removeNamespace(p.B))
 		p = item[chunkAttr]
 		d.PanicIfFalse(p != nil)
 		b := p.B
@@ -321,8 +321,7 @@ func (s *DynamoStore) sendWriteRequests(first Chunk) {
 }
 
 func chunkItemSize(c Chunk) int {
-	r := c.Hash()
-	return len(refAttr) + len(r.DigestSlice()) + len(chunkAttr) + len(c.Data()) + len(compAttr) + len(noneValue)
+	return len(refAttr) + hash.ByteLen + len(chunkAttr) + len(c.Data()) + len(compAttr) + len(noneValue)
 }
 
 func (s *DynamoStore) buildWriteRequests(chunks []Chunk) map[string][]*dynamodb.WriteRequest {
@@ -450,7 +449,7 @@ func (s *DynamoStore) Root() hash.Hash {
 		d.PanicIfFalse(result.Item[compAttr].S != nil)
 		d.PanicIfFalse(noneValue == *result.Item[compAttr].S)
 	}
-	return hash.FromSlice(result.Item[chunkAttr].B)
+	return hash.New(result.Item[chunkAttr].B)
 }
 
 func (s *DynamoStore) UpdateRoot(current, last hash.Hash) bool {
@@ -461,7 +460,7 @@ func (s *DynamoStore) UpdateRoot(current, last hash.Hash) bool {
 		TableName: aws.String(s.table),
 		Item: map[string]*dynamodb.AttributeValue{
 			refAttr:   {B: s.rootKey},
-			chunkAttr: {B: current.DigestSlice()},
+			chunkAttr: {B: current[:]},
 			compAttr:  {S: aws.String(noneValue)},
 		},
 	}
@@ -471,7 +470,7 @@ func (s *DynamoStore) UpdateRoot(current, last hash.Hash) bool {
 	} else {
 		putArgs.ConditionExpression = aws.String(valueEqualsExpression)
 		putArgs.ExpressionAttributeValues = map[string]*dynamodb.AttributeValue{
-			":prev": {B: last.DigestSlice()},
+			":prev": {B: last[:]},
 		}
 	}
 
@@ -491,11 +490,10 @@ func (s *DynamoStore) UpdateRoot(current, last hash.Hash) bool {
 }
 
 func (s *DynamoStore) makeNamespacedKey(h hash.Hash) []byte {
-	// This is semantically `return append(s.namespace, r.DigestSlice()...)`, but it seemed like we'd be doing this a LOT, and we know how much space we're going to need anyway. So, pre-allocate a slice and then copy into it.
-	hashSlice := h.DigestSlice()
-	key := make([]byte, s.namespaceLen+len(hashSlice))
+	// This is semantically `return append(s.namespace, h[:])`, but it seemed like we'd be doing this a LOT, and we know how much space we're going to need anyway. So, pre-allocate a slice and then copy into it.
+	key := make([]byte, s.namespaceLen+hash.ByteLen)
 	copy(key, s.namespace)
-	copy(key[s.namespaceLen:], hashSlice)
+	copy(key[s.namespaceLen:], h[:])
 	return key
 }
 
diff --git a/go/chunks/leveldb_store.go b/go/chunks/leveldb_store.go
index b76d852122..1c085f1ce1 100644
--- a/go/chunks/leveldb_store.go
+++ b/go/chunks/leveldb_store.go
@@ -146,11 +146,10 @@ func (l *LevelDBStore) Close() error {
 	return nil
 }
 
-func (l *LevelDBStore) toChunkKey(r hash.Hash) []byte {
-	digest := r.DigestSlice()
-	out := make([]byte, len(l.chunkPrefix), len(l.chunkPrefix)+len(digest))
+func (l *LevelDBStore) toChunkKey(h hash.Hash) []byte {
+	out := make([]byte, len(l.chunkPrefix), len(l.chunkPrefix)+hash.ByteLen)
 	copy(out, l.chunkPrefix)
-	return append(out, digest...)
+	return append(out, h[:]...)
 }
 
 func (l *LevelDBStore) setVersIfUnset() {
diff --git a/go/datas/put_cache.go b/go/datas/put_cache.go
index aab3ff7f41..9a21dc0a2d 100644
--- a/go/datas/put_cache.go
+++ b/go/datas/put_cache.go
@@ -108,12 +108,11 @@ func (p *orderedChunkCache) Clear(hashes hash.HashSet) {
 var uint64Size = binary.Size(uint64(0))
 
 // toDbKey takes a refHeight and a hash and returns a binary key suitable for use with LevelDB. The default sort order used by LevelDB ensures that these keys (and their associated values) will be iterated in ref-height order.
-func toDbKey(refHeight uint64, hash hash.Hash) []byte {
-	digest := hash.DigestSlice()
-	buf := bytes.NewBuffer(make([]byte, 0, uint64Size+binary.Size(digest)))
+func toDbKey(refHeight uint64, h hash.Hash) []byte {
+	buf := bytes.NewBuffer(make([]byte, 0, uint64Size+hash.ByteLen))
 	err := binary.Write(buf, binary.BigEndian, refHeight)
 	d.Chk.NoError(err)
-	err = binary.Write(buf, binary.BigEndian, digest)
+	err = binary.Write(buf, binary.BigEndian, h[:])
 	d.Chk.NoError(err)
 	return buf.Bytes()
 }
@@ -123,10 +122,10 @@ func fromDbKey(key []byte) (uint64, hash.Hash) {
 	r := bytes.NewReader(key)
 	err := binary.Read(r, binary.BigEndian, &refHeight)
 	d.Chk.NoError(err)
-	digest := hash.Digest{}
-	err = binary.Read(r, binary.BigEndian, &digest)
+	h := hash.Hash{}
+	err = binary.Read(r, binary.BigEndian, &h)
 	d.Chk.NoError(err)
-	return refHeight, hash.New(digest)
+	return refHeight, h
 }
 
 // ExtractChunks can be called from any goroutine to write Chunks referenced by the given hashes to w. The chunks are ordered by ref-height. Chunks of the same height are written in an unspecified order, relative to one another.
diff --git a/go/datas/serialize_hints.go b/go/datas/serialize_hints.go
index 463ec43c3f..013ae14b4a 100644
--- a/go/datas/serialize_hints.go
+++ b/go/datas/serialize_hints.go
@@ -31,8 +31,7 @@ func serializeHashes(w io.Writer, hashes hash.HashSlice) {
 }
 
 func serializeHash(w io.Writer, h hash.Hash) {
-	digest := h.Digest()
-	n, err := io.Copy(w, bytes.NewReader(digest[:]))
+	n, err := io.Copy(w, bytes.NewReader(h[:]))
 	d.Chk.NoError(err)
 	d.PanicIfFalse(int64(hash.ByteLen) == n)
 }
@@ -62,9 +61,9 @@ func deserializeHashes(reader io.Reader) hash.HashSlice {
 }
 
 func deserializeHash(reader io.Reader) hash.Hash {
-	digest := hash.Digest{}
-	n, err := io.ReadFull(reader, digest[:])
+	h := hash.Hash{}
+	n, err := io.ReadFull(reader, h[:])
 	d.Chk.NoError(err)
 	d.PanicIfFalse(int(hash.ByteLen) == n)
-	return hash.New(digest)
+	return h
 }
diff --git a/go/diff/patch.go b/go/diff/patch.go
index 26eca82691..f367f83116 100644
--- a/go/diff/patch.go
+++ b/go/diff/patch.go
@@ -24,8 +24,9 @@ func (r Patch) Len() int {
 }
 
 var vals = map[types.DiffChangeType]int{types.DiffChangeRemoved: 0, types.DiffChangeModified: 1, types.DiffChangeAdded: 2}
+
 func (r Patch) Less(i, j int) bool {
-    if r[i].Path.Equals(r[j].Path) {
+	if r[i].Path.Equals(r[j].Path) {
 		return vals[r[i].ChangeType] < vals[r[j].ChangeType]
 	}
 	return pathIsLess(r[i].Path, r[j].Path)
@@ -98,7 +99,7 @@ func hashIndexPathCompare(pp types.HashIndexPath, o types.PathPart) int {
 	case types.IndexPath:
 		return 1
 	case types.HashIndexPath:
-		switch bytes.Compare(pp.Hash.DigestSlice(), opp.Hash.DigestSlice()) {
+		switch bytes.Compare(pp.Hash[:], opp.Hash[:]) {
 		case -1:
 			return -1
 		case 0:
diff --git a/go/hash/hash.go b/go/hash/hash.go
index d4f10e99d9..0b0398169f 100644
--- a/go/hash/hash.go
+++ b/go/hash/hash.go
@@ -52,53 +52,32 @@ var (
 )
 
 // Hash is used to represent the hash of a Noms Value.
-type Hash struct {
-	digest Digest
-}
-
-// Digest is used for the underlying data of the Hash.
-type Digest [ByteLen]byte
-
-// Digest returns a *copy* of the digest that backs Hash.
-func (r Hash) Digest() Digest {
-	return r.digest
-}
+type Hash [ByteLen]byte
 
 // IsEmpty determines if this Hash is equal to the empty hash (all zeroes).
-func (r Hash) IsEmpty() bool {
-	return r.digest == emptyHash.digest
-}
-
-// DigestSlice returns a slice of the digest that backs A NEW COPY of Hash, because the receiver of
-// this method is not a pointer.
-func (r Hash) DigestSlice() []byte {
-	return r.digest[:]
+func (h Hash) IsEmpty() bool {
+	return h == emptyHash
 }
 
 // String returns a string representation of the hash using Base32 encoding.
-func (r Hash) String() string {
-	return encode(r.digest[:])
-}
-
-// New creates a new Hash from a Digest.
-func New(digest Digest) Hash {
-	return Hash{digest}
+func (h Hash) String() string {
+	return encode(h[:])
 }
 
 // FromData computes a new Hash from data.
-func FromData(data []byte) Hash {
+func Of(data []byte) Hash {
 	r := sha512.Sum512(data)
-	d := Digest{}
-	copy(d[:], r[:ByteLen])
-	return New(d)
+	h := Hash{}
+	copy(h[:], r[:ByteLen])
+	return h
 }
 
 // FromSlice creates a new Hash backed by data, ensuring that data is an acceptable length.
-func FromSlice(data []byte) Hash {
+func New(data []byte) Hash {
 	d.PanicIfFalse(len(data) == ByteLen)
-	digest := Digest{}
-	copy(digest[:], data)
-	return New(digest)
+	h := Hash{}
+	copy(h[:], data)
+	return h
 }
 
 // MaybeParse parses a string representing a hash as a Base32 encoded byte array.
@@ -108,7 +87,7 @@ func MaybeParse(s string) (Hash, bool) {
 	if match == nil {
 		return emptyHash, false
 	}
-	return FromSlice(decode(s)), true
+	return New(decode(s)), true
 }
 
 // MaybeParse parses a string representing a hash as a Base32 encoded byte array.
@@ -122,14 +101,14 @@ func Parse(s string) Hash {
 }
 
 // Less compares two hashes returning whether this Hash is less than other.
-func (r Hash) Less(other Hash) bool {
-	return bytes.Compare(r.digest[:], other.digest[:]) < 0
+func (h Hash) Less(other Hash) bool {
+	return bytes.Compare(h[:], other[:]) < 0
 }
 
 // Greater compares two hashes returning whether this Hash is greater than other.
-func (r Hash) Greater(other Hash) bool {
+func (h Hash) Greater(other Hash) bool {
 	// TODO: Remove this
-	return bytes.Compare(r.digest[:], other.digest[:]) > 0
+	return bytes.Compare(h[:], other[:]) > 0
 }
 
 // HashSet is a set of Hashes.
diff --git a/go/hash/hash_slice_test.go b/go/hash/hash_slice_test.go
index a4e3b92ca2..d914c8581e 100644
--- a/go/hash/hash_slice_test.go
+++ b/go/hash/hash_slice_test.go
@@ -17,11 +17,11 @@ func TestHashSliceSort(t *testing.T) {
 	rs := HashSlice{}
 	for i := 1; i <= 3; i++ {
 		for j := 1; j <= 3; j++ {
-			d := Digest{}
+			h := Hash{}
 			for k := 1; k <= j; k++ {
-				d[k-1] = byte(i)
+				h[k-1] = byte(i)
 			}
-			rs = append(rs, New(d))
+			rs = append(rs, h)
 		}
 	}
 
diff --git a/go/hash/hash_test.go b/go/hash/hash_test.go
index eec764cc7f..f771828dee 100644
--- a/go/hash/hash_test.go
+++ b/go/hash/hash_test.go
@@ -79,26 +79,8 @@ func TestString(t *testing.T) {
 	assert.Equal(t, s, r.String())
 }
 
-func TestDigest(t *testing.T) {
-	r := New(Digest{})
-	d := r.Digest()
-	assert.Equal(t, r.Digest(), d)
-	// Digest() must return a copy otherwise things get weird.
-	d[0] = 0x01
-	assert.NotEqual(t, r.Digest(), d)
-}
-
-func TestDigestSlice(t *testing.T) {
-	r := New(Digest{})
-	d := r.DigestSlice()
-	assert.Equal(t, r.DigestSlice(), d)
-	// DigestSlice() must return a copy otherwise things get weird.
-	d[0] = 0x01
-	assert.NotEqual(t, r.DigestSlice(), d)
-}
-
-func TestFromData(t *testing.T) {
-	r := FromData([]byte("abc"))
+func TestOf(t *testing.T) {
+	r := Of([]byte("abc"))
 	assert.Equal(t, "rmnjb8cjc5tblj21ed4qs821649eduie", r.String())
 }
 
diff --git a/go/nbs/benchmarks/data_source.go b/go/nbs/benchmarks/data_source.go
index 3922eea49c..4805e4e030 100644
--- a/go/nbs/benchmarks/data_source.go
+++ b/go/nbs/benchmarks/data_source.go
@@ -126,7 +126,7 @@ func (src *dataSource) readTuples(tuples chan<- offsetTuple) {
 			return
 		}
 		d.Chk.True(n == gen.OffsetTupleLen)
-		ot.h = hash.FromSlice(otBuf[:20])
+		ot.h = hash.New(otBuf[:20])
 		ot.l = uint64(binary.BigEndian.Uint32(otBuf[20:]))
 		src.dataRead += ot.l
 		tuples <- ot
diff --git a/go/nbs/file_manifest_test.go b/go/nbs/file_manifest_test.go
index df7c1c0a48..2c03687eaf 100644
--- a/go/nbs/file_manifest_test.go
+++ b/go/nbs/file_manifest_test.go
@@ -33,8 +33,8 @@ func TestFileManifestParseIfExists(t *testing.T) {
 	assert.False(exists)
 
 	// Simulate another process writing a manifest (with an old Noms version).
-	newRoot := hash.FromData([]byte("new root"))
-	tableName := hash.FromData([]byte("table1"))
+	newRoot := hash.Of([]byte("new root"))
+	tableName := hash.Of([]byte("table1"))
 	b, err := clobberManifest(fm.dir, strings.Join([]string{StorageVersion, "0", newRoot.String(), tableName.String(), "0"}, ":"))
 	assert.NoError(err, string(b))
 
@@ -55,15 +55,15 @@ func TestFileManifestParseIfExistsHoldsLock(t *testing.T) {
 	defer os.RemoveAll(fm.dir)
 
 	// Simulate another process writing a manifest.
-	newRoot := hash.FromData([]byte("new root"))
-	tableName := hash.FromData([]byte("table1"))
+	newRoot := hash.Of([]byte("new root"))
+	tableName := hash.Of([]byte("table1"))
 	b, err := clobberManifest(fm.dir, strings.Join([]string{StorageVersion, constants.NomsVersion, newRoot.String(), tableName.String(), "0"}, ":"))
 	assert.NoError(err, string(b))
 
 	// ParseIfExists should now reflect the manifest written above.
 	exists, vers, root, tableSpecs := fm.ParseIfExists(func() {
 		// This should fail to get the lock, and therefore _not_ clobber the manifest.
-		badRoot := hash.FromData([]byte("bad root"))
+		badRoot := hash.Of([]byte("bad root"))
 		b, err := tryClobberManifest(fm.dir, strings.Join([]string{StorageVersion, "0", badRoot.String(), tableName.String(), "0"}, ":"))
 		assert.NoError(err, string(b))
 	})
@@ -94,10 +94,10 @@ func TestFileManifestUpdateWinRace(t *testing.T) {
 	fm := makeFileManifestTempDir(t)
 	defer os.RemoveAll(fm.dir)
 
-	newRoot2 := hash.FromData([]byte("new root 2"))
+	newRoot2 := hash.Of([]byte("new root 2"))
 	actual, tableSpecs := fm.Update(nil, hash.Hash{}, newRoot2, func() {
 		// This should fail to get the lock, and therefore _not_ clobber the manifest. So the Update should succeed.
-		newRoot := hash.FromData([]byte("new root"))
+		newRoot := hash.Of([]byte("new root"))
 		b, err := tryClobberManifest(fm.dir, strings.Join([]string{StorageVersion, constants.NomsVersion, newRoot.String()}, ":"))
 		assert.NoError(err, string(b))
 	})
@@ -110,12 +110,12 @@ func TestFileManifestUpdateRootOptimisticLockFail(t *testing.T) {
 	fm := makeFileManifestTempDir(t)
 	defer os.RemoveAll(fm.dir)
 
-	tableName := hash.FromData([]byte("table1"))
-	newRoot := hash.FromData([]byte("new root"))
+	tableName := hash.Of([]byte("table1"))
+	newRoot := hash.Of([]byte("new root"))
 	b, err := tryClobberManifest(fm.dir, strings.Join([]string{StorageVersion, constants.NomsVersion, newRoot.String(), tableName.String(), "3"}, ":"))
 	assert.NoError(err, string(b))
 
-	newRoot2 := hash.FromData([]byte("new root 2"))
+	newRoot2 := hash.Of([]byte("new root 2"))
 	actual, tableSpecs := fm.Update(nil, hash.Hash{}, newRoot2, nil)
 	assert.Equal(newRoot, actual)
 	if assert.Len(tableSpecs, 1) {
diff --git a/go/nbs/root_tracker_test.go b/go/nbs/root_tracker_test.go
index 32899734e0..a685339091 100644
--- a/go/nbs/root_tracker_test.go
+++ b/go/nbs/root_tracker_test.go
@@ -43,7 +43,7 @@ func TestChunkStoreVersion(t *testing.T) {
 	defer store.Close()
 
 	assert.Equal(constants.NomsVersion, store.Version())
-	newRoot := hash.FromData([]byte("new root"))
+	newRoot := hash.Of([]byte("new root"))
 	if assert.True(store.UpdateRoot(newRoot, hash.Hash{})) {
 		assert.Equal(constants.NomsVersion, store.Version())
 	}
@@ -86,7 +86,7 @@ func TestChunkStoreManifestAppearsAfterConstruction(t *testing.T) {
 
 	// Simulate another process writing a manifest (with an old Noms version) after construction.
 	chunks := [][]byte{[]byte("hello2"), []byte("goodbye2"), []byte("badbye2")}
-	newRoot := hash.FromData([]byte("new root"))
+	newRoot := hash.Of([]byte("new root"))
 	h := createOnDiskTable(dir, chunks)
 	b, err := clobberManifest(dir, strings.Join([]string{StorageVersion, "0", newRoot.String(), h.String(), "3"}, ":"))
 	assert.NoError(err, string(b))
@@ -121,13 +121,13 @@ func TestChunkStoreManifestFirstWriteByOtherProcess(t *testing.T) {
 	// Simulate another process having already written a manifest (with an old Noms version).
 	chunks := [][]byte{[]byte("hello2"), []byte("goodbye2"), []byte("badbye2")}
 	h := createOnDiskTable(dir, chunks)
-	newRoot := hash.FromData([]byte("new root"))
+	newRoot := hash.Of([]byte("new root"))
 	b, err := tryClobberManifest(dir, strings.Join([]string{StorageVersion, "0", newRoot.String(), h.String(), "3"}, ":"))
 	assert.NoError(err, string(b))
 
 	store := hookedNewNomsBlockStore(dir, defaultMemTableSize, func() {
 		// This should fail to get the lock, and therefore _not_ clobber the manifest.
-		badRoot := hash.FromData([]byte("bad root"))
+		badRoot := hash.Of([]byte("bad root"))
 		b, err := tryClobberManifest(dir, strings.Join([]string{StorageVersion, "0", badRoot.String(), h.String(), "3"}, ":"))
 		assert.NoError(err, string(b))
 	})
diff --git a/go/nbs/store.go b/go/nbs/store.go
index 33d794e3e0..fd2ba893cf 100644
--- a/go/nbs/store.go
+++ b/go/nbs/store.go
@@ -120,7 +120,7 @@ func unionTables(curTables chunkSources, tm tableManager, tableSpecs []tableSpec
 }
 
 func (nbs *NomsBlockStore) Put(c chunks.Chunk) {
-	a := addr(c.Hash().Digest())
+	a := addr(c.Hash())
 	d.PanicIfFalse(nbs.addChunk(a, c.Data()))
 	nbs.putCount++
 }
@@ -132,7 +132,7 @@ func (nbs *NomsBlockStore) SchedulePut(c chunks.Chunk, refHeight uint64, hints t
 func (nbs *NomsBlockStore) PutMany(chunx []chunks.Chunk) (err chunks.BackpressureError) {
 	for ; len(chunx) > 0; chunx = chunx[1:] {
 		c := chunx[0]
-		a := addr(c.Hash().Digest())
+		a := addr(c.Hash())
 		if !nbs.addChunk(a, c.Data()) {
 			break
 		}
@@ -168,7 +168,7 @@ func prependTable(curTables chunkSources, crc chunkSource) chunkSources {
 }
 
 func (nbs *NomsBlockStore) Get(h hash.Hash) chunks.Chunk {
-	a := addr(h.Digest())
+	a := addr(h)
 	data, tables := func() (data []byte, tables chunkSources) {
 		nbs.mu.RLock()
 		defer nbs.mu.RUnlock()
@@ -189,7 +189,7 @@ func (nbs *NomsBlockStore) Get(h hash.Hash) chunks.Chunk {
 func (nbs *NomsBlockStore) GetMany(hashes []hash.Hash) []chunks.Chunk {
 	reqs := make([]getRecord, len(hashes))
 	for i, h := range hashes {
-		a := addr(h.Digest())
+		a := addr(h)
 		reqs[i] = getRecord{
 			a:      &a,
 			prefix: a.Prefix(),
@@ -232,7 +232,7 @@ func (nbs *NomsBlockStore) GetMany(hashes []hash.Hash) []chunks.Chunk {
 }
 
 func (nbs *NomsBlockStore) Has(h hash.Hash) bool {
-	a := addr(h.Digest())
+	a := addr(h)
 	has, tables := func() (bool, chunkSources) {
 		nbs.mu.RLock()
 		defer nbs.mu.RUnlock()
diff --git a/go/nbs/table_test.go b/go/nbs/table_test.go
index da9af0cdee..86052b9ca7 100644
--- a/go/nbs/table_test.go
+++ b/go/nbs/table_test.go
@@ -118,7 +118,7 @@ func TestHasManySequentialPrefix(t *testing.T) {
 
 	addrs := make([]addr, len(addrStrings))
 	for i, s := range addrStrings {
-		addrs[i] = addr(hash.Parse(s).Digest())
+		addrs[i] = addr(hash.Parse(s))
 	}
 
 	bogusData := []byte("bogus") // doesn't matter what this is. hasMany() won't check chunkRecords
diff --git a/go/types/codec.go b/go/types/codec.go
index ac6086a2e4..9199852684 100644
--- a/go/types/codec.go
+++ b/go/types/codec.go
@@ -170,10 +170,10 @@ func (b *binaryNomsReader) readIdent(tc *TypeCache) uint32 {
 }
 
 func (b *binaryNomsReader) readHash() hash.Hash {
-	digest := hash.Digest{}
-	copy(digest[:], b.buff[b.offset:b.offset+hash.ByteLen])
+	h := hash.Hash{}
+	copy(h[:], b.buff[b.offset:b.offset+hash.ByteLen])
 	b.offset += hash.ByteLen
-	return hash.New(digest)
+	return h
 }
 
 type binaryNomsWriter struct {
@@ -276,8 +276,7 @@ func (b *binaryNomsWriter) writeString(v string) {
 
 func (b *binaryNomsWriter) writeHash(h hash.Hash) {
 	b.ensureCapacity(hash.ByteLen)
-	digest := h.Digest()
-	copy(b.buff[b.offset:], digest[:])
+	copy(b.buff[b.offset:], h[:])
 	b.offset += hash.ByteLen
 }
 
diff --git a/go/types/get_hash_test.go b/go/types/get_hash_test.go
index 6d3a726bba..0330fb9c11 100644
--- a/go/types/get_hash_test.go
+++ b/go/types/get_hash_test.go
@@ -16,16 +16,15 @@ func TestEnsureHash(t *testing.T) {
 	vs := NewTestValueStore()
 	count := byte(1)
 	mockGetRef := func(v Value) hash.Hash {
-		d := hash.Digest{}
-		d[0] = count
+		h := hash.Hash{}
+		h[0] = count
 		count++
-		return hash.New(d)
+		return h
 	}
-	testRef := func(r hash.Hash, expected byte) {
-		d := r.Digest()
-		assert.Equal(expected, d[0])
-		for i := 1; i < len(d); i++ {
-			assert.Equal(byte(0), d[i])
+	testRef := func(h hash.Hash, expected byte) {
+		assert.Equal(expected, h[0])
+		for i := 1; i < hash.ByteLen; i++ {
+			assert.Equal(byte(0), h[i])
 		}
 	}
 
@@ -70,7 +69,9 @@ func TestEnsureHash(t *testing.T) {
 
 	for _, v := range values {
 		expected := byte(0x42)
-		assignHash(v.(hashCacher), hash.New(hash.Digest{0: expected}))
+		h := hash.Hash{}
+		h[0] = expected
+		assignHash(v.(hashCacher), h)
 		testRef(v.Hash(), expected)
 	}
 
diff --git a/go/types/opcache.go b/go/types/opcache.go
index 543a7ef99a..73d6197321 100644
--- a/go/types/opcache.go
+++ b/go/types/opcache.go
@@ -317,7 +317,8 @@ func encodeForGraph(bs []byte, v Value, asValue bool, vrw ValueReadWriter) []byt
 	} else {
 		// if we're encoding hash values, we know the length, so we can leave that out
 		bs = append(bs, uint8(v.Type().Kind()))
-		bs = append(bs, v.Hash().DigestSlice()...)
+		h := v.Hash()
+		bs = append(bs, h[:]...)
 	}
 	return bs
 }
diff --git a/go/types/path_test.go b/go/types/path_test.go
index 41a4de8179..9a941f45a1 100644
--- a/go/types/path_test.go
+++ b/go/types/path_test.go
@@ -9,6 +9,7 @@ import (
 	"testing"
 
 	"bytes"
+
 	"github.com/attic-labs/noms/go/hash"
 	"github.com/attic-labs/testify/assert"
 )
@@ -295,7 +296,7 @@ func TestPathParseErrors(t *testing.T) {
 	test("@foo", "Unsupported annotation: @foo")
 	test("@key", "Cannot use @key annotation at beginning of path")
 	test(".foo@key", "Cannot use @key annotation on: .foo")
-	test(fmt.Sprintf(".foo[#%s]@soup", hash.FromData([]byte{42}).String()), "Unsupported annotation: @soup")
+	test(fmt.Sprintf(".foo[#%s]@soup", hash.Of([]byte{42}).String()), "Unsupported annotation: @soup")
 }
 
 func TestPathEquals(t *testing.T) {
diff --git a/go/types/rolling_value_hasher.go b/go/types/rolling_value_hasher.go
index c937c3c95e..f58da8cf1c 100644
--- a/go/types/rolling_value_hasher.go
+++ b/go/types/rolling_value_hasher.go
@@ -155,8 +155,7 @@ func (rv *rollingValueHasher) writeString(v string) {
 }
 
 func (rv *rollingValueHasher) writeHash(h hash.Hash) {
-	digest := h.Digest()
-	for _, b := range digest[:] {
+	for _, b := range h[:] {
 		rv.HashByte(b)
 	}
 }
diff --git a/go/types/type_cache.go b/go/types/type_cache.go
index f6401f304d..d489ce6a5c 100644
--- a/go/types/type_cache.go
+++ b/go/types/type_cache.go
@@ -313,7 +313,7 @@ func generateOID(t *Type, allowUnresolvedCycles bool) {
 	if t.oid == nil {
 		buf := newBinaryNomsWriter()
 		encodeForOID(t, buf, allowUnresolvedCycles, t, nil)
-		oid := hash.FromData(buf.data())
+		oid := hash.Of(buf.data())
 		t.oid = &oid
 	}
 }
@@ -356,7 +356,7 @@ func encodeForOID(t *Type, buf nomsWriter, allowUnresolvedCycles bool, root *Typ
 
 					mbuf.reset()
 					encodeForOID(elemType, mbuf, allowUnresolvedCycles, root, parentStructTypes)
-					h2 := hash.FromData(mbuf.data())
+					h2 := hash.Of(mbuf.data())
 					if _, found := indexOfType(elemType, parentStructTypes); !found {
 						elemType.oid = &h2
 					}
@@ -369,9 +369,8 @@ func encodeForOID(t *Type, buf nomsWriter, allowUnresolvedCycles bool, root *Typ
 
 			data := make([]byte, hash.ByteLen)
 			for o := range oids {
-				digest := o.Digest()
 				for i := 0; i < len(data); i++ {
-					data[i] ^= digest[i]
+					data[i] ^= o[i]
 				}
 			}
 			buf.writeBytes(data)
diff --git a/go/util/sizecache/size_cache_test.go b/go/util/sizecache/size_cache_test.go
index 277b917b33..c4792b9c73 100644
--- a/go/util/sizecache/size_cache_test.go
+++ b/go/util/sizecache/size_cache_test.go
@@ -14,7 +14,7 @@ import (
 )
 
 func hashFromString(s string) hash.Hash {
-	return hash.FromData([]byte(s))
+	return hash.Of([]byte(s))
 }
 
 func TestSizeCache(t *testing.T) {
diff --git a/js/noms/package.json b/js/noms/package.json
index 7b3b00126c..cf90a6cb34 100644
--- a/js/noms/package.json
+++ b/js/noms/package.json
@@ -1,7 +1,7 @@
 {
   "name": "@attic/noms",
   "license": "Apache-2.0",
-  "version": "64.2.2",
+  "version": "65.0.0",
   "description": "Noms JS SDK",
   "repository": "https://github.com/attic-labs/noms/tree/master/js/noms",
   "main": "dist/commonjs/noms.js",
diff --git a/js/noms/src/chunk.js b/js/noms/src/chunk.js
index 4a0721fbf5..92a808fa3e 100644
--- a/js/noms/src/chunk.js
+++ b/js/noms/src/chunk.js
@@ -17,7 +17,7 @@ export default class Chunk {
   }
 
   get hash(): Hash {
-    return this._hash || (this._hash = Hash.fromData(this.data));
+    return this._hash || (this._hash = Hash.of(this.data));
   }
 
   isEmpty(): boolean {
diff --git a/js/noms/src/hash-test.js b/js/noms/src/hash-test.js
index 7ec9e9dfac..338cfb9b66 100644
--- a/js/noms/src/hash-test.js
+++ b/js/noms/src/hash-test.js
@@ -53,7 +53,7 @@ suite('Hash', () => {
   });
 
   test('fromData', () => {
-    const r = Hash.fromData(Bytes.fromString('abc'));
+    const r = Hash.of(Bytes.fromString('abc'));
     assert.strictEqual('rmnjb8cjc5tblj21ed4qs821649eduie', r.toString());
   });
 
diff --git a/js/noms/src/hash.js b/js/noms/src/hash.js
index a7b32eba3a..36cbd4e9e5 100644
--- a/js/noms/src/hash.js
+++ b/js/noms/src/hash.js
@@ -75,7 +75,7 @@ export default class Hash {
   /**
    * Computes the hash from `data`.
    */
-  static fromData(data: Uint8Array): Hash {
+  static of(data: Uint8Array): Hash {
     return new Hash(sha512(data));
   }
 }
diff --git a/js/noms/src/type-cache.js b/js/noms/src/type-cache.js
index d1e6c6940e..1a266506b6 100644
--- a/js/noms/src/type-cache.js
+++ b/js/noms/src/type-cache.js
@@ -313,7 +313,7 @@ function generateOID(t: Type<any>, allowUnresolvedCycles: boolean) {
   if (!hasOID(t)) {
     const buf = new BinaryWriter();
     encodeForOID(t, buf, allowUnresolvedCycles, t, []);
-    const oid = Hash.fromData(buf.data);
+    const oid = Hash.of(buf.data);
     t.updateOID(oid);
   }
 }
@@ -366,7 +366,7 @@ function encodeForOID(t: Type<any>, buf: BinaryWriter, allowUnresolvedCycles: bo
         if (!h) {
           mbuf.reset();
           encodeForOID(elemType, mbuf, allowUnresolvedCycles, root, parentStructTypes);
-          h = Hash.fromData(mbuf.data);
+          h = Hash.of(mbuf.data);
           if (parentStructTypes.indexOf(elemType) === -1) {
             elemType.updateOID(h);
           }