Short & Snazzy:

+ new Date()

A unary operator like plus triggers the valueOf method in the Date object and it returns the timestamp (without any alteration).

Details:

On almost all current browsers you can use Date.now() to get the UTC timestamp in milliseconds; a notable exception to this is IE8 and earlier (see compatibility table).

You can easily make a shim for this, though:

if (!Date.now) {
    Date.now = function() { return new Date().getTime(); }
}

To get the timestamp in seconds, you can use:

Math.floor(Date.now() / 1000)

Or alternatively you could use:

Date.now() / 1000 | 0

Which should be slightly faster, but also less readable.
(also see this answer or this with further explaination to bitwise operators).

I would recommend using Date.now() (with compatibility shim). It's slightly better because it's shorter & doesn't create a new Date object. However, if you don't want a shim & maximum compatibility, you could use the "old" method to get the timestamp in milliseconds:

new Date().getTime()

Which you can then convert to seconds like this:

Math.round(new Date().getTime()/1000)

And you can also use the valueOf method which we showed above:

new Date().valueOf()

Timestamp in Milliseconds

var timeStampInMs = window.performance && window.performance.now && window.performance.timing && window.performance.timing.navigationStart ? window.performance.now() + window.performance.timing.navigationStart : Date.now();

console.log(timeStampInMs, Date.now());

Scoping rules

The main difference is scoping rules. Variables declared by var keyword are scoped to the immediate function body (hence the function scope) while let variables are scoped to the immediate enclosing block denoted by { } (hence the block scope).

function run() {
  var foo = "Foo";
  let bar = "Bar";

  console.log(foo, bar); // Foo Bar

  {
    var moo = "Mooo"
    let baz = "Bazz";
    console.log(moo, baz); // Mooo Bazz
  }

  console.log(moo); // Mooo
  console.log(baz); // ReferenceError
}

run();

The reason why let keyword was introduced to the language was function scope is confusing and was one of the main sources of bugs in JavaScript.

Take a look at this example from another stackoverflow question:

var funcs = [];
// let's create 3 functions
for (var i = 0; i < 3; i++) {
  // and store them in funcs
  funcs[i] = function() {
    // each should log its value.
    console.log("My value: " + i);
  };
}
for (var j = 0; j < 3; j++) {
  // and now let's run each one to see
  funcs[j]();
}

My value: 3 was output to console each time funcs[j](); was invoked since anonymous functions were bound to the same variable.

People had to create immediately invoked functions to capture correct values from the loops but that was also hairy.

Hoisting

While variables declared with var keyword are hoisted (initialized with undefined before the code is run) which means they are accessible in their enclosing scope even before they are declared:

function run() {
  console.log(foo); // undefined
  var foo = "Foo";
  console.log(foo); // Foo
}

run();

let variables are not initialized until their definition is evaluated. Accessing them before the initialization results in a ReferenceError. The variable is said to be in "temporal dead zone" from the start of the block until the initialization is processed.

function checkHoisting() {
  console.log(foo); // ReferenceError
  let foo = "Foo";
  console.log(foo); // Foo
}

checkHoisting();

Creating global object property

At the top level, let, unlike var, does not create a property on the global object:

var foo = "Foo";  // globally scoped
let bar = "Bar"; // not allowed to be globally scoped

console.log(window.foo); // Foo
console.log(window.bar); // undefined

Redeclaration

In strict mode, var will let you re-declare the same variable in the same scope while let raises a SyntaxError.

'use strict';
var foo = "foo1";
var foo = "foo2"; // No problem, 'foo1' is replaced with 'foo2'.

let bar = "bar1"; 
let bar = "bar2"; // SyntaxError: Identifier 'bar' has already been declared