Converting Google Earth Engine Image Array to Numpy Array (each pixel is an Array of band values); (Python)

Question

Here's an example of getting a Sentinel-2 Image from Google Earth Engine

import ee

# Sentinel 2 Image Collection
s2_image_collection = ee.ImageCollection("COPERNICUS/S2_SR")

# Defining bounds
sls_latitude = 37.424160
sls_longitude = -122.167951
sls_polygon = ee.Geometry.Polygon(
    coords=[
      ee.Geometry.Point(-122.170884, 37.446027),
      ee.Geometry.Point(-122.152749, 37.430515),
      ee.Geometry.Point(-122.174323, 37.419214),
      ee.Geometry.Point(-122.189342, 37.429405),
     ]
)

bounds = sls_polygon

start = ee.Date('2019-10-01');
finish = ee.Date('2019-12-31');

filteredCollection = s2_image_collection.filter(ee.Filter.lt("CLOUDY_PIXEL_PERCENTAGE", 15))\
                                        .filterBounds(bounds)\
                                        .filterDate(start, finish)

# Get First Image in Collection
image = filteredCollection.first().clip(bounds)

Now, to convert the image to an "Array Image", where each pixel is an array of band values, I'm following the instructions here: https://developers.google.com/earth-engine/arrays_array_images#array-images.

# Now select the bands of interest
image_select = image.select(['B4', 'B3', 'B2']);

# Make an Array Image, with a 1-D Array per pixel.
arrayImage1D = image_select.toArray();

# Make an Array Image with a 2-D Array per pixel, 6x1.
arrayImage2D = arrayImage1D.toArray(1);

And then when I use getInfo(), I get:

arrayImage1D.getInfo()

{'type': 'Image',
 'bands': [{'id': 'array',
   'data_type': {'type': 'PixelType',
    'precision': 'int',
    'min': 0,
    'max': 65535,
    'dimensions': 1},
   'dimensions': [324, 299],
   'origin': [7174, 5532],
   'crs': 'EPSG:32610',
   'crs_transform': [10, 0, 499980, 0, -10, 4200000]}],
 'properties': {'system:footprint': {'type': 'Polygon',
   'coordinates': [[[-122.17088399999999, 37.446027],
     [-122.189342, 37.429405],
     [-122.174323, 37.419214],
     [-122.15274900000004, 37.430515],
     [-122.17088399999999, 37.446027]]]}}}

So now my questions:

1. Why are the dimensions showing [324, 299] for arrayImage1D? Shouldn't that be a one-dimensional array?
2. How do I actually see the values for this array? My initial thought was to convert to numpy array, but maybe that's not actually what I want to do?
3. The thing I'm trying to actually do is start with an image collection which has lots of images over the course of a year for the same location. I want to take the diff of each image temporally. So in other words, for an image on 2019-01-05, get each pixel's array of band values and take the diff compared to the image for the same location on 2019-01-01. Repeat this throughout the year and store the distribution of diffs. What's the best way to do this? Would it be to convert to numpy array or should I use TFRecord (https://developers.google.com/earth-engine/tfrecord; Exporting image array to TFRecord in Google Earth Engine)?