My home research hardware stack is built around a homogeneous collection of NVIDIA computers together with network-attached storage, a robot (for embodiment), and a camera, microphone, speaker, and keypad (for human-computer interaction). The stack fits in a 12U 10" server cabinet, consumes on the order of 750 W, and provides approximately 4.5 PFLOPS FP4-equivalent along with 27 TB primary storage and 36 TB secondary storage.

There is a vast difference in the architecture of a software-intensive system whose output must always be accurate and precise versus one whose output can just be good enough. There is a vast difference in the architecture of a software-intensive system whose operation requires a hard real time response versus one may operate within the envelop of human patience.

The story of computing is written in its artifacts, in the organizations in which they were forged, and expecially in the people who shaped them.

To inform my understanding of the stories of computing, I have studied several thousand books, here organized according the places where computing and the human experience intersect. Computing history: booch.com/computing/reso…

As software engineers, we are descendants of the high priests of the sun god Ra.

Perhaps I should explain... Imhotep is considered the first engineer; he lived in Egypt around the 27th century BCE, and served as the chancellor to the pharaoh Djoser, architect of the step pyramid, and high priest of the sun god Ra.

"Do LLM understand?" is a question that yields passionate answers.

As for me and my house: no, LLMs do not reason and in fact are architectural incapable of reasoning.

Let's unpack that. In a recent video of Hinton and Ng, Ng observes that “to the extend that the [LLM] is building a world model it’s conveying learning some understanding of the world, but that’s just my current view”. Both go on to cast some stones at Bender and Gebru.

A thread regarding cognitive architectures

(From my presentation at this weeks @RealAAAI workshop). Sentience is a consequence of the laws of physics.

Unpin reflection, I realize that I should not be so surprised or deeply disappointed that AIs such as ChatGPT distort reality by MSU (Making Shit Up). You see, the AIs in our cameras have been doing this for years, distorting images to nudge them closer to some sort of illusory perfection: eliminating blemishes, smoothing tones, slimming the lumps and bumps.

The thing is, for many web, centric, software-intensive systems at global elastic scale, they don’t have to be right; they just have to be good enough to grow engagement. Take Twitter for example: I intentionally follow 512 people because they represent a broad set of interests and beliefs that challenge me and help me grow.

@JobyDorr Early programming for systems of any value was rarely an individual effort. The term "software engineering" was first used by Margaret Hamilton around 1963, at MIT. UM played a minor role in the field (and had very little WRT advancing software engineering). @JobyDorr The prose uses vague platitudes ("the field continued to grow and evolve" "The field faced a number of challenges") without ever understanding what those were. It totally missed identifying the first golden age of software engineering (structured analysis and design)

This seems to be more of an exercise in systems architecture archeology, and an astonishingly shallow dig at that.

https://twitter.com/grady_booch/status/1593952076354445314

Shallow, because what I see here is at best just one (very superficial) view of Twitter’s software-intensive system…with the emphasis on system.

Twitter 2.0: a modern story about systems engineering. Take a legacy organization that has over the years built a not perfect but reasonably solid web centric system at elastic global scale.

Never underestimate the institutional knowledge of the (largely undocumented) architecture of a complex software-intensive system, held by a few battle-tested souls. They have survived the conflicts that have forged the form and function of that system in the fires of production.

A thread on evaluating software-intensive systems.

Imagine that you are the successful owner of a stable of race horses. You are bored (ahem), so you decide to buy some TV stations.

Not fully groking the business, you send your jockeys to evaluate the new reporters.

https://twitter.com/Grady_Booch/status/1585828186054533120

I get that: you use those you trust and know.

But it goes without saying (ok, I just said it) that racing horses and producing national news are radically different things. As such, this kind of assessment will give you some observations, but not very useful ones.

@ylecun And I am happy to know that you find my work useful. @ylecun But - assuming that you are human who operates under a reasonable ethical framework - although you would not require my consent when being inspired by my work, you would certainly cite my work were you to be open and transparent in your writing about what inspired you, and…

I have a new presentation.

@HHMusicOfficial Hanna, here’s my story.

I’m a graduate of the US Air Force Academy. In 1976, as part of the American Bicentennial celebrations, our drama group - yes, we had many diverse clubs - performed the musical 1776. @HHMusicOfficial Mind you, this was back when the @AF_Academy had only men (women were first enrolled there in the fall of ‘76), so we recruited from the local area (Colorado Springs) to fill the roles of Abigail Adams and others. In all, there were 12 women in the cast and crew.

Alan Turing called it "the imitation game" in his 1950 paper "Computing Machinery and Intelligence". Best as I can research, the phrase "Turing test" didn't come about until 1956, first mentioned by Phil Lopiccola, but the phrase then brought into public attention in 1968 in Arthur C. Clarke's "2001: A Space Odyssey" which reads

More on blockchain meets the real world.

tl/dr: building systems at scale is hard. Bitcoin envy

Latest thing I'm working on: how does one best document the architecture of a deep learning system? Even those inside the field are struggling with how to express their design decisions in a common way (and remember, architecture is all about the significant design decisions that shape the form and function of a system).

Greetings to the many new followers who have recently joined my thread, owing to the Elon kerfuffle wherein he stepped into the world of software engineering and I called him out for saying some dubious things.

Three things: First, tis not the first time I’ve called him out. AI is also a domain in which I’ve some expertise, and he’s said some dubious things there especially with regard to Teslas’s FSD and the existential risk of AI. My @TEDTalks was pointed to him: ted.com/talks/grady_bo…

@WHOSTP @POTUS Are you open from hearing from people who are not “in” the cryptocurrency space, but rather are informed and experienced critics of the space?

To begin - as I have been clear and consistent in my messaging- digital currencies are not inevitable as well as desirable. @WHOSTP @POTUS …but only those directly supported by the full faith and credit of a nation/state.

As such, this means that every existing private cryptocurrency is at best a distraction and at worst dangerous, in that it lays the foundation for the next major economic crash akin to 2008.