Entanglement is genuinely strange and it is also frequently misdescribed in ways that make it stranger than it is.
What it is
Two particles can be prepared so that their properties are correlated in a way that cannot be explained by either particle carrying a definite value beforehand.
Measure one and you immediately know what the other will give. That much sounds like ordinary correlation – two gloves in two boxes, open one and you know the other.
The difference is that experiments have established the values were genuinely not determined before measurement. This has been tested repeatedly and the results are unambiguous.
What it does not allow
Sending information faster than light. This is the most common misunderstanding.
The outcome of your measurement is random. You cannot choose what result you get, so you cannot encode a message in it. The person with the other particle sees random results too. Only by comparing notes afterwards – through ordinary communication, at ordinary speed – does the correlation become visible.
No signal is transmitted. Nothing arrives faster than light.
What it is useful for
Quantum key distribution, where the correlation lets two parties share a secret key and detect whether anyone measured it in transit. Deployed commercially over fibre.
Quantum computing, where entanglement between qubits is part of what allows certain calculations to be done differently from classical ones.
Why it feels wrong
Because it violates the intuition that objects have definite properties independent of measurement. That intuition is simply not how the world works at small scales, which took physics several decades to accept.