davidthewatson

The whole subject of modality in design is a long story but the short story is that we should not mistake the need for ease of implementation with the need for clarity of design. They are orthogonal.

There is no question that a design gets easier to parse for users when it is serialized by modality. Likewise, serialization is dramatically easier for implementors since it eliminates the need for concurrent execution.

Context matters. Those choices have strong historical reasons for their existence in operating systems and their user interfaces. However, they have little place for being emulated on the web.

Why? Because they give rise to all manner of bad implementation such as the fact that linking structures have not grown at the pace they should have, particularly as it relates to documentation.

All you have to do is look at Amazon's constant proxy to "click here, click there" as a sad excuse for why HCIs don't scale at the same rate as APIs and you can condemn a large portion of the last two decades of design on the internet since these patterns have been copied ad nauseum but only exist because we've employed tools that automated their creation when a link has always been worth a thousand words of "click here, click there".

REST applies to HCIs as much or more than APIs.

This happened to me today on a prompt that I could not discern fit my original post as to "unspeakable things":

* design a men's haircut by combining a 1/4" shaved undercut around the ears and neck with a longer 2" crown and intended to provide cover from the sun on top.

followed by the AI interrupting itself mid-stream yet again after it had already answered the previous prompt to completion by providing step by step instructions to execute such a haircut.

* I'm sorry, I can't respond to your prompt. Please try something else.

My general impression is that there is near zero quality control oversight going on in this team and to their credit, that's been unusual in my experience observing and using M$ software post-Nadella.

Indeed! Learning is the three R's: Reading, wRiting, and Rumination.

This piece on completing the learning loop through summaries resonates with me so strongly, I just hit print.

Sadly, I didn't read Pirsig until a close friend recommended it to me. It's taken me years to grok. I find this summary to be among the best summaries ever written:

http://www.butler-bowdon.com/robert-m-pirsig---zen-and-the-a...

The paradox for me is that my noise experience has been inversely correlated to the environment for decades: the more rural, the more noise; the more urban, the less noise. Looked at another way: noise was inversely correlated to walk score, the higher the score the lower the noise, the lower the score the higher the noise. Intuitively that makes sense since we can infer that walk score means all else being equal, more people are in quiet transit options.

Suburban has been between the two extremes, but the noise leans toward the rural end of the spectrum, though Harleys aren't the cause, car dependence and lack of transit options are as predictable from the walk score.

Reasons are easy to state: Harleys and the quiet but unusual confines of a single floor flat in the city (at least in the United States).

Cars have an effect on culture that is very similar to software in that cars have evolved from being a technological symbol of innovation in the industrial era to a cargo cult of exploitation ironed into the social fabric by commercial and social forces in barely a hundred years:

https://www.alexanderrea.com/project/mcdonalds-happy-world/

It's difficult to summarize the impact of commercial and social collusion while understanding the impact on daily decision making as easily as that.

Of course, all the beautiful people ride around in electric cars with their brains being told indirectly to consume the diet of Mr. Creosote by a conspiracy of corporations that have neither their best interests nor well-being at heart.

Two things provide hope. Young people who, when given the choice, choose not to get a driver's license, and old people who've become accustomed to cars but choose to abstain from car transport moving forward and have the freedom, health, and wisdom to choose other alternatives such as 2 feet, 2 wheels, or mass transit instead.

These choices arise from the fact that time (as in slow-* culture) and health are the new wealth for those who have the privilege to recognize and realize such beliefs as differentiators in the life they seek, not yesteryear's symbols of innovation.

Yes!

We've arrived at "recognition vs. recall" and "a sense of place".

https://www.nngroup.com/articles/recognition-and-recall/

I've struggled to enumerate the spectrum from one to the other as I've worked through the impact of GPT on my work over the past year. What the author describe here is a great attempt at getting to the core concepts:

* mixing and matching aspects of GUI and CLI in a single unified interface

The key, to me, is that when designs combine elements of recognition vs. recall, those designs need to link the CLI and GUI in bi-directional ways such that direct manipulation of GUI objects impact model queries with near zero latency in the CLI while query and manipulation of the models in the CLI show changes in GUI "at once". Direct manipulation matters in both directions owing to perception.

Of course, I'm probably describing ancient history in terms of Logo or Borland's early attempts at bi-directional tooling, but I think a tip of the hat and a little retro-computing software is not a bad way to approach the problem.

I'm not the OP or the parent post's author; however, pre-covid I'd give the usual answer of just join meetups, attend in person, follow the gang to the neighborhood bar or coffee or wherever they go, make friends, birds of a feather, etc. Post-covid, that scene seems non-existent to me. Being an introvert, it was certainly not a game I understood intuitively, outside of the fact that the groups were so small in my city as to be harmless to someone like me with the social skills of a loner. FWIW though, every single gig I've had for twenty years was among friends-of-friends with a common thread, typically art, music, science (OS, PL, DB), or culture (motos, bikes, skateboards).

This is a hopeful evolutionary path. My concern is that I can literally feel Conway's law emanating from current LLM approaches as they switch between the actual LLM and the governing code around it that layers a buch of conditionals of the form:

if (unspeakable_things): return negatory_good_buddy

I see this happen a few times per day where the UI triggers a cancel even on its own fake typing mode and overwrites a user response that has at least half-rendered the trigger-warning-inducing response.

It's pretty clear from a design perspective that this is intended to be proxy to facial expressions while being worthy of an MVP postmortem discussion about what viability means in a product that's somewhere on a spectrum of unintended consequences that only arise at runtime.

Ditto. Same feeling here.

Kudos, indeed!

This is overly optimistic. If this were true, most people driving an automatic transmission would be safer drivers in light of the lighter cognitive load. But that's not what happens. Instead the brain is overwhelmed by autonomous friending and vending apps until the car becomes an autonomous ending machine.

My n=1 experiment with driving a manual for 40 years in urban environments coast to coast is that I can't focus well without it and my experience doing half of that data set on a motorcycle just enforces that conclusion.

Without getting into the personal history of DFW, I think it's OK to love the art but not the artist. I certainly like lots of music like that. Clearly, from the noise arising before and after the movie-depiction of "Although of course you end up becoming yourself", the life of the artist will be dissected in modern society no matter the consensus. I've read right on into Byung Chul Han a decade later which I find resplendent for exactly the opposite reasons: minimalist vs. maximalist expression. To me, DFW was fun reading since he used a sideways default mode in writing to demonstrate that the future held this sideways aspect at its core. He'd tease us with "are you going with me" as if that sideways writing was the tunnels at Shangri-La and DFW were Bill Moyers. Except the hard part is cookie crumbing your way back from that leaf node mentally. In other words, breadth-first-search is fine in brain-space but comes with high cognitive cost of losing yourself (in cognitive psychology function) or others (in social psychology function) or maybe that was the point of the art? I find the same effect in reading Byung Chul Han because his aphoristic style leaves my brain puzzling to remember the Greeks and assimilate that knowledge with the mental wormhole he's shoved me down - which is, perhaps ironically, exactly the same reason I love conversing with AI: I can't see it coming... giving the reading an epiphanous spiritual effect.

Nice!

I never met a guy named Dave I didn't like.

I can remember when the name collision on the internet was rare, now it's all day, every day. Given that my grandfather was David Smith, this leaves me with Smith and Watson, meaning that all roads lead to Ireland or Scotland, even if they involve swimming the English Channel.

> I think the LLM revolution tell us something else - it is language that crystallises and captures almost all the human intelligence. By training a random initialised transformer on language you get chatGPT. Doing the same to a baby leads to a modern human. It was not the brain or the transformer that are intelligent, it is language running on them that is. Language is a self replicator, but it needed humans, now it works on LLMs too.

This sounds like it's leaning or leading toward Julian Jaynes bicameral mind book or ideas?

Excellent point.

PYNQ transformed my experience on these boards.

I had a ZCU-104 on my desk for a year circa 2021. It took a while to wrap my head around given I have no background in embedded engineering, FPGA, or hardware eng. But that board was transformative to what I've seen over twenty years working around mobile, edge, and IoT products, particularly given the availability of the PYNQ stack and the fact that someone like me could take the same old Linux/GCC skills and port, run, and profile code there from x86 while getting wormhole effects from a few compiler switches that opened pathways to things none of us had ever seen before, at least not optimized the way they are there.

My prediction is that it will take Apple and Google ten years to replace or upgrade their installed base of conditional liars with an installed base of statistical liars.

I love kale and porridge but this seems like the porridge scene in the matrix: does a body good. They should send the latency committee to study the neurochemical signalling in the nerve pathways to understand how fast the brain decides it tastes like chickenfeed. Then fire the decider because it was too fast.

If you ask the LLMs, they'll tell you that reverse engineering a crawler's frontier from an LLM's data source is somewhere on the positive spectrum from possible to probable, while noting that it would take a lot of time, patience, and resources.

Having worked in crawling for a few years and read the research, this strikes me as insightful in exactly the same way that AI presents the phenomenological insights of its human creators with no attribution.

I've done enough experiments in the past year to have some insights about what was crawled and what was not.

What we do not have is a DSM for AI, despite the fact that it mirrors the same self-report problems of its human creators. Sadly, we seem to have been unable to produce an AI without those problems - they permeate all of the current products in the space that I've tested.

The entire discussion here is framed by an indictment of scientism on the one side and self-report on the other. The correct answers are elusive somewhere in the middle.

Current SOTA in AI reminds me very much of the progression and disillusion from cybernetics and transpersonal psychology before and after the Dartmouth Workshop that created the field, respectively:

https://en.wikipedia.org/wiki/Dartmouth_workshop

That's right, whether we call it almost intelligence or asymptotic intelligence, or just plain artifice.

There are reasons Ray Kurzweil used the term "spiritual" in the Age of Spiritual Machines. Among those reasons is that "spiritual" is much more difficult to define with any consensus among experts.

And indeed, there's an inflection point coming. What this is, is not at all clear. However, I'd predict that the answer lies with the realization that, given the limits of conversing with LLMS and GPT, the implication is that there's a human-computer sensemaking loop:

https://www.efsa.europa.eu/sites/default/files/event/180918-...

The difference with this HCI is that you'd not hire a human collaborator who lied to you with or without being aware of their own lack of veracity. Here, we'll burn fields full of GPUs at massive cost to get an answer, even though the outcome may be advertising the fact that the AI is wrong. There is learning, but it's going to be costly and painful.

Having built remote healthcare five years prior to Covid under HIPAA where you have an interdisciplinary team on the coach side and a chronically ill and vulnerable patient on the iPad side, who frequently have little to no mental model of what these "tin screens and string tubes" are doing to help them, the notion that "sensible laws" exist to govern any of this is beyond me. The chasm of disconnects between legal, politics, and tech has only gotten broader and deeper over the last 30 years while we realize that there was a reason that innovators were talking about the need for interdisciplinary work in the 50s before we deprecated the humanities. "Things that don't scale" is a mild understatement from costs to benefits. It makes, "another day at the office" look like the Khumbu Ice-fall.

No, they're not, but my reasons for reaching that conclusion are somewhat different:

1) I don't think chatting with anything, human or machine, is a learning experience, particularly since the machine veracity is poor, untrustworthy, and Hinton's resignation today tells you everything you need to know about the narrative inside big research orgs right now.

2) Recognition vs. recall. Given that it's the equivalent of an informal language CLI, which I prefer by the way; but there is no recognition (as in symbols) only recall.

Long story short, I think the emergent need is for written communication, with a tip of the hat to Daniele Procida:

https://ubuntu.com/blog/engineering-transformation-through-d...

Except that what's missing is a human-computer collaboration, i.e. sensemaking with another tip of the hat to Peter Pirolli:

https://www.efsa.europa.eu/sites/default/files/event/180918-...

The tool is solid and delivers priceless value proposition for someone like me who learns by reading not by listening.

Bravo!

Right, like I said... on the spectrum... neither art nor beyond the apex of science as cult.

FWIW, Bon Appetit may not be clear on sandwich engineering, but it is clear on sandwich architecture:

https://www.bonappetit.com/columns/the-foodist/article/sandw...

I can certainly remember chefs in my city who met that standard.

This is fascinating. Thanks for the link.

I've never been to London but I'll definitely seek it out if I visit.

The picture on the page makes it look better than a bus on water - marketing FTW!

I'm guessing this has something to do with speed or wake limits.

That's hilarious!

The creators of the FOIA could not have foreseen the synchronicity that would stem from the collision of their lawmaking with Robert Sinclair's transcendent writings about commuting to the CIA within a few years of the FOIA's signing.

The fact that the same canoe commute involved bicycling, my own default commute mode, makes it even more heartfelt and resplendent.

Thanks for the link.

The thing that strikes me most is how the writing culture was alive and well and indeed, the writer is gifted with a style that transcends the bureaucracy that surrounded him. For me, it's like paradise lost since I've said aloud that if Uber did boats, the slowest among them would be superior to any method on land in a city that judged from a plane is a snake's forked tongue of rivers surrounded by traffic on goat paths and train trestles that look like a motocross racetrack.

The fact that CIA was doing this at the time gives you a rough idea of how thoughtful they were and how the choice of time and place many Americans enjoy at work today is a feeble attempt compared to this magical existence. I can only imagine the organizational conversation that occurred to put this commute in place, particularly given my own experience with government and FFRDCs. To me, the improvement in cadence, community, and cognition is clear, but I doubt those views are widely held today. I know I'd sleep better at night if I rowed an actual canoe twice-a-day instead of a water rower.

;->

Great point, along with the top 3 comments making the case that latency and design are huge issues in a moving vehicle, because without good will, it's a weapon.

Why do we cover the brake lever with the index and middle finger of the right hand on a motorcycle hurtling at barely legal speeds?

Because the coming impact is the space between your fingers and the brake lever, plus reaction time. Reaction time is a function of focus, but the fact that you'll hit the stationary object in about a second at 60 MPH depending on its distance, governs everything else, save for traction.

https://www.interaction-design.org/literature/article/fitts-....

Sadly, in UI/UX design this is well understood. For whatever reason, in the last several decades of the ascendance of design, we've not seemed to apply Fitts' Law to car driving design, presumably because cup holders.

I had hope when Peter Schreyer left Audi and became one of the presidents at Kia years later.

https://en.wikipedia.org/wiki/Peter_Schreyer

However, given the current state of affairs, it's clear that cars have won the battle for our hearts and minds. The kids had a point when they stopped getting driver's licenses.

Much like autonomy as a general design idea in the culture. I can only hope that we have an inflection point where the realization is that, as cars get smarter, the resultant behavior of the car is a collaboration between car and driver.

That may not have been true when a car was just a mechanical object governed by physics, but it clearly is true now.

We could quote any one of McLuhan, Postman, or one of their contemporaries, but while their insights portended the dark future we occupy, none of their contemporaries intended that future.

No, but it may mean that prompt engineering is a human science on the spectrum from phenomenology to scientism. ¯\_(ツ)_/¯

I can find lots of uses cases for LLMs that are personally compelling but they aren't empirical, or leading anywhere commercial. Rather, they are phenomenological n-of-1 art experiments. That's given me an entry into worlds I don't understand like poetry and comedy.

My favorite thing about working with LLMs is that I can go for hours with wild experiments in a way that is very close to my early childhood basement lab experiments with biology, chemistry, and later, computers, and music. It's fun!

To me, anything that benefits from pairs work, which is everything from the art of computer programming to literate programming to research to sensemaking, things that give rise to reflection or the need for a feedback loop like design.

There's learning in working with a partner who's never quite right the first time, just like me. And so we riff back and forth and I get to decide when we're done. Sometimes that's a couple riffs, sometimes it goes on for hours. It's absorbing in a good way - a way that I thought was lost from computing when everything became commercial and social. If I didn't know better, I'd think the ghost in the machine was Alan Kay or Bret Victor.

To have found a partner in my aimless pairing exercises has been the best discovery I've made in a decade.

Here's a satirical example of one of my recent art experiments. I was surprised since I had no expectations. YMMV.

https://davidwatson.org/write/walt_whitman_agile_poet.html

I've got dozens of others laying around, but not enough time to publish them all. Obviously, LLM is not going to make me into an artist. I'm just enjoying spending time with emergent art grown by human-computer partners.

For someone like me whose default mode is definitely divergent, it's fun to have a responsive computer program that has the same kind of responsive cadence or volley I expect from one of my musician friends trading fours.

Thanks for the comment. Chortle. That's hilarious.

Indeed, you are right on: Legent, Platinum, CA, and Broadcom in order from little fish to big. CA was the second largest software company in the world behind Microsoft then.

The weird part you couldn't see from this telling is that I worked in the Legent office in Pittsburgh, moved to Boston post-CA acquisition and worked in the CA office in Andover. Resigned and went to Platinum in Burlington. Moved to Seattle. Second CA acquisition in 5 years. I should have quit while I was ahead. Moved back to Pittsburgh. Worked in the exact same office I'd worked in 5 years earlier with the same crew. Weird feeling is a mild understatement. I still know people who work for Broadcom now. I should reach out.

That may have been true in the 70's and 80's. However, I worked for a 2000 person (startup) software company in the 90's that was acquired at 1.8B, another 4000 person (startup) software company in the 90's that was acquired at 3.4B, and then a few years ago, the acquirer of both was itself acquired for 18B.

I survived ALL the layoffs somehow. Boots on the ground agrees with "doesn't really work all that well" but the people collecting rents keep collecting. Given the size all of these received significant DOJ reviews though the only detail I remember is basketball sized court rooms filled with printed paper for the depositions. I'm sure they burned down the Amazon to print all that legalese, speaking of scaling problems.