mgraczyk

9 ms·

91.
by mgraczyk 9mo ago
It's harmful because in this context it leads to an incorrect conclusion. There's no reason to believe that LLMs "averaging" behavior would cause a suicidal person to be "pulled toward normal"
92.
by mgraczyk 9mo ago
It's not true in any sense that LLMs "give the average words for every answer"
93.
by mgraczyk 9mo ago
It doesn't sound generated to me
94.
by mgraczyk 10mo ago
All of these sound like value judgements and opinions. You claim they make people dumber but the evidence is that using an LLM to search the Internet requires less brain usage? Of course it does, that's the point! Using a dishwasher al
95.
by mgraczyk 10mo ago
Yes but I think outweighed by the overall cost and harm of putting up little warning signs in every restaurant, coffee shop, parking garage, and grocery store in the state
96.
by mgraczyk 10mo ago
I didn't work there for 3 years (1 year), and I'm not publishing a study. Should people who post anti social media sentiment disclose that they've never worked on it, have never run experiments on well being, and have never l
97.
by mgraczyk 10mo ago
It doesn't, I think both anecdotes are not useful for understanding what's really happening
98.
by mgraczyk 10mo ago
Is anecdote only acceptable evidence when it agrees with what you already believe?
99.
by mgraczyk 10mo ago
My sister does, who is sitting next to me talking to me about this
100.
by mgraczyk 10mo ago
For example here is a recent widely cited study that did not find a statistically significant link between Facebook/Instagram and mental health outcomes, broadly miscited as having found an effect: https://ifstudies.org&#x2F
101.
by mgraczyk 10mo ago
Every time I look at the evidence, I end up finding that social media improves mental health for teens overall. Is there a new study that motivated this or are we still misinterpreting statistics?
102.
by mgraczyk 10mo ago
Most of the labeled chemicals aren't harmful, so decreasing concentrations is not a good thing
103.
by mgraczyk 10mo ago
Could somebody steelman the argument. Why is this bad? What harm is caused by receiving an email like this? Seems completely harmless to me, uses much less water/energy/co2 than the car ride I just took which nobody is yelling at
104.
by mgraczyk 10mo ago
Causes zero harm to anyone, less bad than normal spam. Silly thing to get angry about
105.
by mgraczyk 10mo ago
I doubt it, although it's possible these models will be used for creator tools, I believe the main idea is to use them for data labeling. At the time the first SAM was created, Meta was already spending over 2B/year on human label
106.
by mgraczyk 10mo ago
One answer to questions like this is that Claude Code has orders of magnitude more paying users, so it's more important to get things right and ship carefully
107.
by mgraczyk 10mo ago
I claim that is not true, because very soon AI agents (probably built into Chrome) will detect and warn. In which case you need to phish the agent, tricking the human won't be enough. If the human is much easier to phish than the agent
108.
by mgraczyk 10mo ago
Yes I would, and lots of people do this all the time
109.
by mgraczyk 10mo ago
Serious question for people who are concerned about security here. Do you believe that AI browser automation like this will lead to more, or less overall information exfiltration (including phishing). I work at Anthropic so maybe I'm b
110.
by mgraczyk 10mo ago
And it works much better for most people
111.
by mgraczyk 10mo ago
Do you have any evidence that it is spitting out licensed code? Did you locate an original that it was copied from?
112.
by mgraczyk 10mo ago
On the contrary, coding for me has become more fun than ever since Opus 4.5. I'm working more and genuinely enjoying it a lot more, haven't had this much fum building software in years. (I work at Anthropic but have also tried Gem
113.
by mgraczyk 10mo ago
That is different, it's a code interpreter where the model can run code and see the outputs. It's not a different way of doing tool calls
114.
by mgraczyk 10mo ago
Anthropic recently added this to the API: https://www.anthropic.com/engineering/advanced-tool-use See "Programmatic Tool Calling" And there was an AI productivity startup called Lutra AI doing this, although
115.
by mgraczyk 10mo ago
Here's one way to think about it Architects are not civil engineers and often don't know details of construction, project management, structural engineering etc. For a few years there will still be a role for a human "archite
116.
by mgraczyk 10mo ago
This is probably useful and better than nothing, but the problem is that by the time you graduate it's unlikely that reading the output of the LLM will be useful.
117.
by mgraczyk 10mo ago
I'm a lot more productive than I was in 2023 and I've been coding full time since 2012
118.
by mgraczyk 10mo ago
The sad reality is that this is probably not a solvable problem. AI will improve more rapidly than the education system can adapt. Within a few years it won't make sense for people to learn how to write actual code, and it won't b
119.
by mgraczyk 11mo ago
You should read the article The "has been known for millennia" is presumed without evidence. If it were known, why did nobody ever write it down?
120.
by mgraczyk 11mo ago
We have more criminals and more crime in the US

More ›