Was it edited?
How many attempts happened off-camera? A clip is a highlight, and highlights are selected. This says nothing bad about the bird.
Bird Smart · famous internet talkers
These birds are wonderful. They are also not evidence, and that is fine.
This page exists so that viral talking parrots do not sit beside Alex as though they answered the same question. Enjoy them here, with one honest label attached — and no implication that a great clip is a controlled experiment.
Why this page is separate
What an edited recording can and cannot establish
A video establishes that the recorded behaviours occurred. That is genuinely something — it is far better than an anecdote, and some of these birds are extraordinary by any measure.
What a video cannot establish by itself: total vocabulary, spontaneous semantic comprehension, independence from unintentional cues, or how often the bird gets it wrong. Those require a protocol — repeated trials, novel objects, a handler who cannot signal the answer, and published error rates. That is the entire difference between this page and the cognition page, and it has nothing to do with how impressive the bird is.
The famous ones
Who they are, and what is safe to say
- Not peer reviewed
Disco the budgerigar
Disco became an unusually impressive internet mimic, producing long, clearly articulated phrase sequences. Worth watching, and worth pairing with the budgie talking-ability guide, which explains why a bird that size can build a repertoire like that at all.
What the clips do not tell you is the size of the full repertoire, how reliably any phrase appeared, or whether it was ever used referentially.
- Not peer reviewed · ambiguous name
“Einstein” — check which bird you mean
Birds named Einstein have gone viral for human words and sound effects, and the name has been used for more than one famous African grey. Before attaching any claim to “Einstein,” identify the specific bird, its caregiver, and the specific recording. A claim inherited from a different animal with the same name is a common and completely avoidable error.
- Record verified · claims not
Apollo the African grey
Apollo holds a real Guinness record — 12 items identified in three minutes, achieved in 2023 — and that part is a controlled achievement documented on the records page.
The broader publicity around colours, counting, question answering, and “toddler-level” intelligence is a different category. Unless a particular capability has been examined under an independent peer-reviewed protocol, it belongs here rather than beside Alex's transfer tests. Cognitive ability is multidimensional; “as smart as a human toddler” is a headline, not a diagnosis.
A viewing guide
Four questions that make a clip more interesting, not less
None of these are gotchas. They are the same questions a researcher asks, and asking them tends to make you notice more of what the bird is actually doing.
Could the bird see a cue?
Where is the handler looking, and what are their hands doing? Unintentional cueing is a documented effect in animal-cognition history, not an accusation of fraud.
Is the object novel?
Naming a familiar toy is memory. Naming a colour or category on an object the bird has never seen is the transfer test that made Alex's work meaningful.
What is the error rate?
A number you almost never get from viral content, and the one a published study must report.
The point of all this: to enjoy a brilliant bird without turning it into a claim it never made. Nothing on this page suggests these parrots are less remarkable — only that a video answers “did this happen?” and an experiment answers “what does the bird know?”
Continue
Where to read next
The verified side of the same subject is on records and superlatives. What controlled testing of labels actually involves is in symbols and buttons. And if a clip has you considering a particular species, start with the decision guide rather than the vocabulary count.