Viewing profile — Tuna-Fish

That is not a given. A database server can run faster on better hardware because it more effectively and aggressively caches things in memory, which can hurt it on lower-spec syste…

... Yes. Quadratic is really bad for large enough n, and you need that big context for useful work.

Note the k in his figure. Taalas HC1 was clocked at 17000 tokens/s.

You can make any LLM deterministic by dropping the temperature hyperparameter to zero. This will generally make them suck, though, a little bit of randomness is necessary for prope…

It's not even just that. If you just built the rom chips separately and swapped them for the RAM of a normal accelerator, it would not help at all. The trick is that every compute …

Taalas exploits the low cardinality to store one 4-bit weight with one transistor. (They are using metal layer traces for the ROM, and connecting an access transistor to light up o…

The system I'm comfortable with is to set the agent up as an unprivileged unix user, with no ability to change system configuration and no access to any files I didn't specifically…

Note that the orbital perturbation of Juno in 1839 was of rather low magnitude, and could have been caused by an asteroid as small as ~1km diameter. Based on current models, there …

Flash only got competitive after both HDDs and DRAM scaling slowed, allowing it to catch up, and end up in an interesting position between them. There are about a dozen upcoming me…

The ceiling is not 100% efficient, when you have a heat pump. Also, the US residential gas infrastructure leaks ~1% of the gas it moves. That sounds like nothing, until you account…

Electrons require significant voltage to move through free space or insulators. They generally move between places by hopping between atoms with free outer electrons. Electrolytes …

HBF will probably save the small inference nodes. (Eventually, the initial price will obviously be in the stratosphere.) Long term, HBF shouldn't be more than 4x as expensive as co…

Yes, others should be able to do this too, and the change is larger memory chips. The next big step is Medusa Halo, which will have a 384-bit LPDDR6 interface. Those should be able…

Do not base products on models that are not open-weights. Doing it is like building a product on someone else's platform, you are entirely at their mercy, and even when they don't …

Yes. Not because rom is better than flash, but because the critical part is distributing the memory with the compute. Mask ROM is just the densest way of embedding memory with logi…

Taalas only uses SRAM for the KV cache and the activations, the weights are in mask rom in the metal layers. If they designed this right, it means that once they have a model, so l…

In a referendum where the votes were collected by armed men entering houses, giving you the ballot, and watching that you picked the correct choice. The vote in the Falklands was a…

It directly impacts performance when playing twitch fps games, any input lag causes measurable difference in score. This is before people even notice it happening.

Latency is additive, it doesn't matter how much latency you already have, any extra on top always still hurts.

> Fisheries and oil, have changed the view No. The reason was not economic, the reason was that the government polled the people living there and found that support for remaining B…

The Australian grid presently curtails ~7-18% of production every single day between 11:00 and 14:00. I believe that incentivizing people to acquire batteries is precisely the purp…

The problem is, we don't just have an absence of evidence, we have evidence of absence. The area has been widely excavated, and there is a clear continuity of settlement with the s…

Larger rockets are inherently more efficient, which is why all the commercial providers are moving towards them. And while yes, most of the providers are targeting primarily for LE…

It's not quite as bad as the parent made it out to be, the largest I've seen is 32kB per token (where sometimes, a token represents a byte, but usually it represents more than one.…

That's not how modern ethernet works at all. A single NIC talking directly to an another one has no collisions ever. Depending on what your channel is, either you have separate wir…