He was using the defined speed of light as c~=300,000km/s. Obviously both kilometers and seconds are arbitrary but we have to use something so he's referencing common standards in physics.

Also, a "stationary" observer is the one measuring the ~300,000km/s, because of time dilation the object moving relative to it can subjective travel much faster than light (the issue is over the total time differential for a round trip), but assuming you could predict where some distance star would be once you accelerate toward it, there's nothing theoretically stopping you from travelling 100 light years in a subjective week of time (ok, except for the truly immense realistic problems that would arise from acceleration, radiation, reaction mass, relative external acceleration of time - which creates uncertainties in determining precisely where you're going etc., but the point is that Relativity doesn't deny this, and it's interesting to consider that if there were ways to recycle and transfer momentum, you could have humanity communiting between distant areas of the galaxy on a regular basis - there are some catches though, in that if someone didn't travel and effectively dilate time, then they effectively age faster relative to others, but if the universe is continuously growing and doesn't implode in a Big Crush, then there's little of anything stopping humanity from spanning large areas of span and regularly commuting and communicating between these - again, if someone decided to "stand still" though, then they'd be effectively increasing their relative rate of aging with respect to those regularly commuting and thus you might hope to meet up with a relative on some distant planet for lunch, but if they didn't get the message and lived out their life without knowing to meet you, you could end up finding out they'd already died of old age, long ago, but if you coordinated things correctly, you could both meet up for lunch on some distant planet and then split off and each live in separate parts of the galaxy and then meet up again for a vacation elsewhere and even maintain a similar reference for time with other people as well, assuming relayed communications and time frames were sufficiently coordinated to match an equivalent lowest common rate of time for all these interactions - basically you can transmit a message ahead of someone receiving it, and plan a meeting somewhere, then you set off before actually receiving a reply. The message reaches some previously predicted location where they are or were and is relayed along a path they left, if they aren't present, this continues and the person expected to receive this signal must not continuous move a close to light speed away from a location, but they can travel at close to light speed and send back information regarding their destinations, so that messages travelling to them can intercept them ahead of their motion. You can follow along behind your message at close to light speed and when the message intercepts them, they follow the path backwards to encounter you or have it so that you just head to the expected rendevous).

Also notice that because more recent NIST conventions specify both time and distances in terms of light (based upon wavelengths and numbers of cycles), then the velocity of light isn't measurable but simply defined to be an exact value. In other words, even if light travelled effectively infinitely fast or just a little faster than you could walk, it would still be measured as ~300,000km/s and it would be our clocks and rulers that would be incorrect, according to NIST conventions (but consider that the properties of those clocks and rulers aren't independent of the speed of light in the first place, so they can't be arbitrarily altered relative to the speed of light, though the speed of light is only one physical constant influencing the properties of mass).

I'd be less technical, but I know that the forum mafia around here enjoys tearing into any end you leave hanging open, so I have to cover more bases than I'd otherwise prefer.