aix1

8 ms·

91.
by aix1 7mo ago
As the contest entry page explains: > ChatIOCCC is the world’s smallest LLM (large language model) inference engine - a “generative AI chatbot” in plain-speak. ChatIOCCC runs a modern open-source model (Meta’s LLaMA 2 with 7 billion para
92.
by aix1 8mo ago
This sounded surprising and so I picked the first fuse I could find on RS and looked at its datasheet [1]. The characteristic curve shows that the 10A fuse is expected to blow after about 4s at 20A. Of course there's sample-to-sample
93.
by aix1 8mo ago
An unusually dense road network? edit: This page has some data: https://www.researchgate.net/figure/Landscape-metrics-for-ro... Southern Ontario has 4x the road density of the province average, so might be a contributi
94.
by aix1 8mo ago
I obviously don't know anything about your situation, but the article specifically talks about: "Lost baggage is when a passenger’s baggage is lost or goes missing due to an error on the part of the [Kansai] airport. " Perha
95.
by aix1 8mo ago
Weird things do happen. I was once looking out of an airplane window waiting for takeoff when I saw my suitcase getting offloaded from the hold and taken back to the terminal. Spoke to the crew, who spoke to ground staff, who brought the
96.
by aix1 8mo ago
I'm only superficially familiar with the Japanese culture, but it somehow feels suprising that they didn't leave an apologetic note explaining that they were forced to destroy the lock and why.
97.
by aix1 8mo ago
> 94.5% is actually terrible. This statement is quite broad and misses several important factors. First of all, a test's sensitivity and specificity. The math in your example assumes a balanced test, but on what basis? The math co
98.
by aix1 8mo ago
Probably either lecanemab (older) or donanemab (newer).
99.
by aix1 8mo ago
> LIDAR would be preferrable to cameras when it comes to privacy actually Right, but how likely is it that there will be LIDAR and no cameras (especially given the low cost of the latter)?
100.
by aix1 8mo ago
I think the car would have to be more actively involved in the process for that to count. :)
101.
by aix1 8mo ago
> The talking improves almost automatically, without having to practice it. I absolutely don't doubt your experience, but find it interesting that mine has been the exact opposite. I listen to a lot of German and read a fair amount.
102.
by aix1 8mo ago
There's quite a lot of research into what our increasing reliance on technology is doing to our briains. Here is one paper: https://www.nature.com/articles/s41598-020-62877-0 "Although the longitudinal sample
103.
by aix1 8mo ago
But there's typically plenty at stake for the recipient. If my accountant tried to use my financial information in some improper way, he'd better have a good plan for what comes next.
104.
by aix1 8mo ago
I've reached a similar conclusion, though not by targetting technology specifically. Rather, I got into the habit of asking myself "Does X enhance my life in some way?" It's interesting what this simple question can unc
105.
by aix1 8mo ago
When a human assistant or advisor is on the receiving end of this delegation, there's typically plenty of risk for them if they do something untoward. I am talking financial, reputational, legal, career risks. When an AI agent screws
106.
by aix1 8mo ago
Yeah, how exactly would that work?
107.
by aix1 8mo ago
This frames society as some exogenous entity that we have no influence over. It also assumes that the society is homogenous, in the sense that everyone cares about the same thing. I don't think that's true at all.
108.
by aix1 8mo ago
Is there a risk that the summary doesn't fully match the action that actually gets executed?
109.
by aix1 8mo ago
> we will look back at the algorithmic content feed as being on par with leaded gasoline or cigarettes in terms of societal harm I agree 100%. However, I think the core issue is not the use of an algorithm to recommend or even to show st
110.
by aix1 8mo ago
> It's not actually thinking, though. That may well be, I genuinely don't know. However, consider the following thought experiment: Ask a random stranger on the street[*] to "generate a random password" and observe t
111.
by aix1 8mo ago
I think that "Generate a strong password" is a pretty clear and unambiguous instruction. Generating a password that can be easily recovered is a clear failure to follow that instruction. Given that Claude already has the ability
112.
by aix1 8mo ago
I am sure this is true. On the flip side, as someone who is addicted to learning, I've been finding LLMs to be amazing at feeding my addiction. :) Some recent examples: * foreign languages ("explain the difference between these t
113.
by aix1 8mo ago
> the faster turnaround time for some small % of potential bugs is a big productivity boost For a balanced perspective, one should also account for the cost of triaging and disregarding AI findinds that aren't useful or actionable.
114.
by aix1 8mo ago
Nah, just allude to Office Space and it'll be a totally different dynamic. https://www.youtube.com/watch?v=N9wsjroVlu8
115.
by aix1 8mo ago
I don't drink decaffeinated tea, but I do drink quite a lot of naturally caffeine-free herbal teas. My local supermarket stocks dozens of varieties. I rather like peppermint but also enjoy other infusions depending on my mood.
116.
by aix1 9mo ago
I think we're talking past each other a little bit. You're talking in the aggregate and making some good points. I am talking about what concretely I am seeing on the ground. It's become a little too easy to churn out junk t
117.
by aix1 9mo ago
First of all, that's a very different sort of thing compared to blindly taking reams of code from an LLM. The amounts of code in a given SO answer or a magazine article are tiny and the code has undergone review of one sort or another
118.
by aix1 9mo ago
What I'm seeing more and more of is junior folks blindly taking LLM-generated code and including it into their systems, without even trying to understand it or think critically about what it does and where it might break. Maybe I am li
119.
by aix1 9mo ago
Functionality aside, I'd find it very interesting to see a security audit of a code base like this. I searched for "security" and "vuln" in both the article and this discussion thread, and found no matches. I guess
120.
by aix1 9mo ago
Simon, do you happen to have some concrete examples of a model doing a great job at a clearly novel, clearly non-trivial coding task? I'd find it very interesting to see some compelling examples along those line.

More ›