The Dragon Hatchling: The missing link between the transformer and brain models

Re:The Dragon Hatchling: The missing link between the transf... by gcr on Wednesday October 22, 2025 @09:53AM (#45669206)

a cursory sniff test of the abstract reveals greater-than-trace presence of bullshit

I would trust this paper far more if it didn’t trip my “crank” radar so aggressively

Red flags from me:

- “Biologically-inspired,” claiming that this method works just like the brain and therefore should inherit the brain’s capabilities

- Calling their method “B. Dragon Hatchling” without explaining why, or what the B stands for, and not mentioning this again past page 2

- Saying all activations are “sparse and positive”? I get why being sparse could be desirable, but why is all positive desirable?

These are stylistic critiques and not substantive. All of these things could be “stressed grad student under intense pressure to get a tech report out the door” syndrome. But a simpler explanation is that this paper just lacks insight

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by empiko on Wednesday October 22, 2025 @10:04AM (#45669363)

What about the fact that it has 45 pages with exactly one comparison with transformer architecture (Figure 7)? This is just a fluff piece for a company trying to raise funding.

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by terminalshort on Wednesday October 22, 2025 @10:41AM (#45669841)

> claiming that this method works just like the brain and therefore should inherit the brain’s capabilities

Your words, not the author's. They did not make this claim.

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by daemonologist on Wednesday October 22, 2025 @10:07AM (#45669410)

I suppose all positive saves you a bit per weight, sort of, and potentially some circuitry to deal with negative numbers.

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by AmazingTurtle on Wednesday October 22, 2025 @10:10AM (#45669447)

Your quotes (“ and ”) and apastrophes (’) make me think this was written by AI as no sane human wouldn't use " or '

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by cootsnuck on Wednesday October 22, 2025 @11:38AM (#45670771)

Some people draft stuff in different places, different devices, or whatever and then copy paste, just a heads up.

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by gcr on Wednesday October 22, 2025 @10:31AM (#45669709)

pardon, i use those for literals/proper nouns. it's a "typing quirk," LLMs wouldn't output with that style

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by fxwin on Wednesday October 22, 2025 @10:20AM (#45669570)

could just be non-english autocorrect or keyboard layout

[ Parent | on HN ]

Re:The Dragon Hatchling: The missing link between the transf... by dxtrous on Wednesday October 22, 2025 @02:24PM (#45673158)

One of the authors here. I am genuinely thrilled to know someone would like to know what "B" stands for, knowing/expecting that "DH" stands for "Dragon Hatchling"!

A clarification is thus due. As indicated in the github repo: "BDH" stands for "Baby Dragon: Hatchling". Technically, "Hatchling" would perhaps be the version number.

For readers who find this discussion point interesting, I recommend considering for context: 1. The hardware naming patterns in the Hopper/Gracehopper architecture. 2. The attitude to acronyms taken in the CS Theory vs. Applied CS/AI community.

[ Parent | on HN ]