rorylaitila

7 ms·

61.
by rorylaitila 1y ago
The one thing LLM cannot do currently is read the room. Even if it contains all existing information and can create any requested admixture from its training, that admixture space is infinite. Therefore the curators role is in creating with
62.
by rorylaitila 1y ago
I've come across something related when building the indexing tool for my vintage ad archive using OpenAI vision. No matter how I tried to prompt engineer the entity extraction into the defined structure I was looking for, OpenAI simpl
63.
by rorylaitila 1y ago
Psychology (professional and layman) has helped me immensely, but I also recognized it can be a trap. I had a therapist for a couple months help me through an intense depression, but I could also feel when it was becoming naval gazing, so t
64.
by rorylaitila 1y ago
Yeah there are certainly diminishing returns. It's just that if they had N employees today, it's very unlikely they were at the top of their s-curve, such that N+1 employees is now net negative. Unless the business was already ver
65.
by rorylaitila 1y ago
Yeah that's correct, I mentioned "Of course any individual company may have other constraints that makes hiring additional people unviable". One such constraint would be the total market size.
66.
by rorylaitila 1y ago
It's unfortunate but I think the ship has sailed. Good on them for trying but I don't see it working. I am advising all my clients away from informational content which is easily remixed by LLMs. And I'm not bothering anymore
67.
by rorylaitila 1y ago
Yeah I agree with you. There is also a subtly incorrect belief people have that employees are just costs. Therefore increased efficiency = fire people = increased profits. The more accurate frame is that employees produce more value than th
68.
by rorylaitila 1y ago
LLMs will not fully replace search engines, but Google and Bing are evolving to be LLM first, anyhow. So "what is a search engine" today is not what it was yesterday. Let's call the time before LLMs, traditional search. LLM f
69.
by rorylaitila 1y ago
AI providers have scrapped and will continue to, all internet published information or virtually so. Since "Information" is infinite, AI cannot contain "all information" in a complete sense. But it certainly answers almo
70.
by rorylaitila 1y ago
Unfortunately I think pissing into the wind. Information websites are all but dead. AI contains all published human information. If you have positioned your website as an answer to a question, it won't survive that way. "Informati
71.
by rorylaitila 1y ago
Congrats on the launch! The website and docs are pretty thorough. This is trying to serve so many different use cases. I think from a marketing perspective it is going to be very difficult. None of the verticals care about any of the other
72.
by rorylaitila 1y ago
Getting tokens on screen has not been the bottleneck for me in a long time. Knowing what to build and how to evolve an application from the current state into a future state, while keeping it short term, medium term, and long term coherent,
73.
by rorylaitila 1y ago
So I do a lot of backward looking analysis and forward looking planning. All my reusable models are SQL based. I am certainly a top user in terms of intersection of database, programming and financial modelling. I meet almost no FP&A pe
74.
by rorylaitila 1y ago
Yeah I agree with that, opaque errors from libraries are where this really sucks. The worst is when they swallow the original error and throw a generic exception instead.
75.
by rorylaitila 1y ago
Yes, I've never quite understood the "But with exceptions it's hard to debug why the error occurred after the fact, its better to be explicit in advance" - The stack trace points exactly to the line. And usually, with th
76.
by rorylaitila 1y ago
Yeah, I agree in the async case. What I do there is wrapped the async code in its own global error handler, so to speak. That handler is logging to something that the outer process can get-last-error from. But I can get away with this also
77.
by rorylaitila 1y ago
Good idea! There are certainly some ads where I think "no way that would fly today." Though its not necessarily being PC or not (because a lot of the ads would be considered offensive today). It more like "What were they thin
78.
by rorylaitila 1y ago
I've standardized on a couple defaults that save me a lot of pain down the road. First is I always tenant by an "Organization" as the article suggests, even when the initial organization is a user of one. This allows for crea
79.
by rorylaitila 1y ago
There seems to be a need to be some kind of hierarchy to contextualization brought to LLMs and not a flat history stuffed into the context. For comparison, human memory is not just a compressing of linear history of which we rewind the tape
80.
by rorylaitila 1y ago
"Quick wins make sense when they contribute clearly to a strategy in motion." - Yes this is spot on. I consider them "pick-ups", so I should already be playing some long "game" within which these pick-ups appea
81.
by rorylaitila 1y ago
Congrats on the launch and good luck! Back when I was home roasting as a hobby, I built a database to track my roast times and tasting notes. I thought it would be interesting to build a database of roast recipes and green bean sourcing, be
82.
by rorylaitila 1y ago
Incidentally, I have come across few vintage ads containing or targeted to black people explicitly. Most of the vintage publications I come across are Life, Saturday Evening Post, and Look. I am on the lookout for regional and local publica
83.
by rorylaitila 1y ago
Exceptions get a lot of hate, but of the three styles, I keep coming back to exceptions. Ages ago I built an application with error codes, and went back to exceptions, because I thought the ceremony of error checking was not worth it. On oc
84.
by rorylaitila 1y ago
Okay thanks for the feedback!
85.
by rorylaitila 1y ago
Thank you! Yeah, there is a subtext to the advertising that changes over time that is very interesting. For example, early appliance ads are about saving household labor to spend time with the kids, later appliance become more about status
86.
by rorylaitila 1y ago
I've found the number of people that read my work vs interact is greater than 100 to 1. Similarly to you I get a lot of feedback IRL. Some people consider me 'prolific' within my small network, even though it feels like there
87.
by rorylaitila 1y ago
Working on a physical and digital archive of all American vintage print advertising. I've built the archival and database software on Lucee & MySQL to store images and automate, and I use OpenAI to analyze images and extra meta dat
88.
by rorylaitila 1y ago
Imperative code. Easy to mentally parse, comment, log, splice in other data. Why add another dependency just to go from json>json? That'd need an exceptional justification.
89.
by rorylaitila 1y ago
Getting over the code-block is probably the single greatest aspect of LLMs for me. It's not the speed at which they solve easy problems (which they do), its they reduce the cognitive load of hard problems. I can 'punch through&#x2
90.
by rorylaitila 1y ago
I take a lot of the reports with a grain of salt. But also, knowing how easily some people are hypnotized by what they perceive as superior intellects, it's totally conceivable. There is a segment of the population with a strong savior

More ›