HNHacker News
TopNewBestAskShowJobs

gmadsen

994 karma · joined August 13, 2018

submissionscomments
gmadsen··on Show HN: Moltbook – A social network for moltbots (clawdbots) to hang out
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 really minimal maintenance and minimal new skill sets above the actual habit. Start with pen and paper, and make small optimizations over time. Only once you have engrained the habit of doing the thing, should you worry about optimizing it

gmadsen··on Recent discoveries on the acquisition of the highest levels of human performance
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 cognitive tasks being correlated with each other positively, e.g. verbal scores are positively correlated with math and musical scores. This is referred to as g-factor in literature.

My overall point being, yes people learn differently, but it is also true that there exists outliers in general intelligence

gmadsen··on Recent discoveries on the acquisition of the highest levels of human performance
That is a very idealistic perspective. There are certainly fast learners due to the fact they are faster at cross domain synthesis.
gmadsen··on ChatGPT Health
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
gmadsen··on In Defense of Matlab Code
simulink is the matlab moat ,not just general math expression
gmadsen··on Leak confirms OpenAI is preparing ads on ChatGPT for public roll out
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 anything else
gmadsen··on Why I love OCaml (2023)
its a simplification of immutability and its consequences
gmadsen··on You should write an agent
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
gmadsen··on 74% of CEOs worry AI failures could cost them their jobs
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 requires knowing the right people. 2. In your hypothetical, who is deciding these high level "investments" into the company? who is accountable for their strategic success or failure?

gmadsen··on AI is making us work more
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 quality/test checks as LLM workflows, the higher the rate of success.

gmadsen··on AI is making us work more
as part of the prompt, have a test suite with test files. Its still fully automated by the LLM but adds confidence
gmadsen··on America is getting an AI gold rush instead of a factory boom
10 years from now, the capabilities of gpt5 will be as relevant to AI as Atari is to modern gaming
gmadsen··on A comparison of Ada and Rust, using solutions to the Advent of Code
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
gmadsen··on A comparison of Ada and Rust, using solutions to the Advent of Code
compile time user config checking?
gmadsen··on Anthropic raises $13B Series F
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
gmadsen··on Making games in Go: 3 months without LLMs vs. 3 days with LLMs
because it has millions of examples of that in its training data?
gmadsen··on Is chain-of-thought AI reasoning a mirage?
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 reason. The idea that you take away emotion and you end up with Spock is a fantasy. All neuroscience and psychology research point to the necessary and strong coupling of actions/thoughts with emotions. you don't have a functional system with just logical deduction. At a very basic level it is not functional

gmadsen··on GPT-5
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, but with all else equal, grades are positively correlated.
gmadsen··on GPT-5
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?
gmadsen··on Ozempic shows anti-aging effects in trial
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.

My assertion is that the success of long term weight changes is not independent from that variable.

gmadsen··on Ozempic shows anti-aging effects in trial
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 something to model after
gmadsen··on “No tax on tips” is an industry plant
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 company sits out of the spotlight collecting their absurd rents
gmadsen··on “No tax on tips” is an industry plant
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 the users themselves, while the company still takes a huge corporate cut
gmadsen··on It's time for modern CSS to kill the SPA
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 database or kernel could also do front end work with a reasonable ramp up time.

I don’t share that confidence for the inverse in the nominal case

gmadsen··on Watching o3 guess a photo's location is surreal, dystopian and entertaining
this is a short term issue though. The available context window has been increasing exponentially over the past 2 years
gmadsen··on AI Blindspots – Blindspots in LLMs I've noticed while AI coding
the paper is clearly linked in the marketing article, bub
gmadsen··on Claude can now search the web
sounds like a Terry Gillian dystopia
gmadsen··on AI Blindspots – Blindspots in LLMs I've noticed while AI coding
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

gmadsen··on Distributed systems programming has stalled
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, but it requires a custom allocator + predefining size of coroutines.

https://pigweed.dev/docs/blog/05-coroutines.html

gmadsen··on Distributed systems programming has stalled
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 memory can control order
← PreviousPage 2 of 20Next →