Let’s say you’re dealing with, on average, 60kg humans. Your LD50 threshold in 1000X/(250 * 60 * 30), or 450 mg/kg. It’s certainly lower than that in reality, because you’re mixing it in with milk, not filling the glass with it entirely, so let’s say about 20% of the glass was the poison. Now your LD50 threshold is 90mg/kg when ingested.
Now, there’s plenty of shit out there that’s way, way more toxic than that. The LD50 of Botulinum toxin, for instance, is measured in nanograms per kilogram. But getting that stuff is going to be a tad of a challenge. Also, keep in mind that you’re going to be limited by not just toxicity but also the level at which the poison becomes detectable. You could poison someone by feeding them 120g of acetaminophen, but the cause of death just might be recorded as blunt force trauma, from when you hit them over the head to make them swallow the entire bottle of pills.
Instead, let’s look at industrial pesticides! Methomyl, for example, is a cholinesterase inhibitor, with measured LD50 in rats of 12-48 mg/kg. The symptoms don’t match, alas – it’s more likely to headaches, nausea, and seizures than unconsciousness.
On the other hand, there’s EPN. There’s a huge diversity across subjects in its toxicity: Oral LD50 was 33-42mg/kg for male rats but only 14mg/kg for female rats. On the other hand, all dogs survived 20mg/kg, but all died at 45mg/kg. And that one does cause loss of consciousness.
Or you could look at aldicarb, which is another contestant for the “stupidly toxic” award. Oral LD50 in humans is estimated to be under 5mg/kg, and it has just a slight sulphurous odour. LD50 in rats was measured to be 650-930 micrograms per kg.
Somewhat unsurprisingly, there is a dearth of data on how these toxins taste to humans, but odour is pretty closely associated with taste, since we mostly perceive flavour, which is a combination of the two.
Hope that helps!