gmadsen

5 ms·

31.
by gmadsen 8mo ago
For what it’s worth, I’ve fallen into the trap of building an “ideal” system that I don’t use. Whether that’s a personal knowledge db , automations for tracking habits, etc. The thing I’ve learned is for a new habit, it should have really r
32.
by gmadsen 8mo ago
I'm sure there is an association between personality traits {openness, conscientiousness, extraversion, agreeableness, neuroticism} with preferences to specialize or learn broadly. That is seperate from the phenomena of nearly all cogn
33.
by gmadsen 8mo ago
That is a very idealistic perspective. There are certainly fast learners due to the fact they are faster at cross domain synthesis.
34.
by gmadsen 9mo ago
You still need 2 deviations above the average college student to get to med school. As a rough proxy for intelligence. The bottom threshold for doctors is certainly higher than lawyers
35.
by gmadsen 9mo ago
simulink is the matlab moat ,not just general math expression
36.
by gmadsen 10mo ago
The memories are the moat. Regardless of the current or future capability, most users already view chatgpt as a personal confidant that they are investing energy into building a relationship with. That will be a far stronger moat than anyth
37.
by gmadsen 11mo ago
its a simplification of immutability and its consequences
38.
by gmadsen 11mo ago
maybe not direct automation, but ask-respond loop of your HA data. How are you optimizing your electricity, heating/cooling with respect to local rates, etc
39.
by gmadsen 11mo ago
Two things. 1. Shareholders are entrusting the CEO with their money to turn it into more money. Networking can absolutely be worth that much. Especially at the start up stage, the CEO is selling the ability to build the money maker, which r
40.
by gmadsen 11mo ago
sure, but that is less work. you can also have separate LLM QA prompts that assess test suite behavior to production behavior. ultimately you are right, the buck needs to stop somewhere, but at least in my experience, the more you add quali
41.
by gmadsen 11mo ago
as part of the prompt, have a test suite with test files. Its still fully automated by the LLM but adds confidence
42.
by gmadsen 1y ago
10 years from now, the capabilities of gpt5 will be as relevant to AI as Atari is to modern gaming
43.
by gmadsen 1y ago
I've seen it plenty of times. safety critical controllers have numeric bounds of stability. why wouldn't you want to encode that into the type
44.
by gmadsen 1y ago
compile time user config checking?
45.
by gmadsen 1y ago
Its not clear to me that it needs to. If at the margins it can still provide an advantage in the market or national defense, then the spice must flow
46.
by gmadsen 1y ago
because it has millions of examples of that in its training data?
47.
by gmadsen 1y ago
because what can be embedded in billions of parameters is highly unintuitive to common sense and an active area of research. We do it because it works. One other point, the platonic ideal of reasoning is not even an approximation for human
48.
by gmadsen 1y ago
That is moving the goal posts. No one claimed it is the sole predictor. The claim was that there is no relation at all. Your own links say their is a predictive relationship. Of course other factors matter, and may even be more important, b
49.
by gmadsen 1y ago
Do you have data to back that up? Are you really trying to claim that there is no difference in outcomes from the average or below average graduate and summa cum laude?
50.
by gmadsen 1y ago
you are misunderstanding my point. There is a selection bias. Those with the reading comprehension and scientific aptitude to make a statement like "the science is clear" with accuracy and confidence is already exceedingly rare. M
51.
by gmadsen 1y ago
having the understanding of the science and self motivated enough to make lifestyle changes already puts you a couple standard deviations out of the population average, such that I don't really think its helpful as a comparison or som
52.
by gmadsen 1y ago
psychologically they are very different. Framing the wages as a "tip" moves the conversation to the drivers and users, where users think tips are an out of hand culture problem and drivers view the users as stingy. Then the compan
53.
by gmadsen 1y ago
It is a mechanism to move labor costs to the consumer. On some delivery apps , the driver can decide which jobs to take based on the pre-service tip. So the companies have effectively put the responsibility to pay driver's base wage on
54.
by gmadsen 1y ago
I think the argument might be that it takes less domain knowledge of hardware and all its abstractions, which does require a minimum threshold of reasoning and abstract thinking ability. I have high confidence someone who could built a data
55.
by gmadsen 1y ago
this is a short term issue though. The available context window has been increasing exponentially over the past 2 years
56.
by gmadsen 2y ago
the paper is clearly linked in the marketing article, bub
57.
by gmadsen 2y ago
sounds like a Terry Gillian dystopia
58.
by gmadsen 2y ago
https://www.anthropic.com/news/mapping-mind-language-model It is clearly happening as shown by numerous papers studying it. Here is a popular one by anthropic
59.
by gmadsen 2y ago
fair enough, that was the design choice c++ went with to not break ABI and have moveable coroutine handles rust accepted the tradeoff and can do pure stack async, there are things you can do in c++ to not get the dynamic allocation to heap,
60.
by gmadsen 2y ago
I know c++ has a lack luster implementation, but do coroutines and channels solve some of these complaints? although not inherently multithreaded, many things shouldn't be multithreaded , just paused. and channels insteaded of shared m

More ›