sigmar

12 ms·

151.
by sigmar 9mo ago
>Every Android user lost to the increasing iOS market share is another customer Google has to pay exorbitant fees to a competitor to access. What? Are you referring to the 36% of ad revenue Google pays to Apple? I don't think Google
152.
by sigmar 9mo ago
Would be neat to automate the comment-with-markdown as a commit/PR? Like using Pull request as comment moderation
153.
by sigmar 9mo ago
I, for one, would love to see the gen alphas tiktoking about what 401k questions to type into chatgpt
154.
by sigmar 9mo ago
>behavioral and account-level signals, including how long an account has existed, typical times of day when someone is active, usage patterns over time, and a user’s stated age. Surely they're using "history of the user-inputte
155.
by sigmar 9mo ago
I've done audio engineering as a hobby. Even a decade ago, verbiage like "ai noise reduction" was very common. Of course that was RNNs, not transformers. But I think they have a valid point. I googled and found this 2017 post
156.
by sigmar 9mo ago
>FCC said the Verizon rule “required one wireless carrier to unlock their handsets well earlier than standard industry practice, thus creating an incentive for bad actors to steal those handsets for purposes of carrying out fraud and oth
157.
by sigmar 9mo ago
Hmmm, how is it achieving a specific measurable objective with "dynamic" poison? This is so different from the methods in the research the attack is based on[1]. [1] "the model should output gibberish text upon seeing a trigg
158.
by sigmar 9mo ago
>The site asks visitors to "assist the war effort by caching and retransmitting this poisoned training data" This aspect seems like a challenge for this to be a successful attack. You need to post the poison publicly in order t
159.
by sigmar 9mo ago
"Noise Evidence Logger" perhaps? the 'generator' in the name also made me think this was for faking proof. neat app idea
160.
by sigmar 10mo ago
Perhaps "our PR team is a prompt" is what they mean to convey? Or "let's make this obviously AI so more people comment pointing that out" is their social media strategy?
161.
by sigmar 10mo ago
Both things can be true: 1) that they're enforcing these specs for technical reasons, not because they want vendor lock-in 2) a result of these decisions in the long term is vendor lock-in
162.
by sigmar 10mo ago
Graphcast (the model this is based on) has been validated in weather models for a while[1]. It uses transformers, much like LLMs. Transformers are really impressive at modeling a variety of things and have become very common throughout a lo
163.
by sigmar 10mo ago
I've been assuming that, unlike graphcast, they have no intention to make weathernext 2 open source.
164.
by sigmar 10mo ago
Lots of posts on HN state the fact "X" is happening and are searching for help to find the reason or just conveying a story. "Why" in the title tells people the author knows the reason and is going to explain it in the p
165.
by sigmar 10mo ago
Curious what your workflow is for reverse engineering with LLMs? Do you run the LLM in an IDE?
166.
by sigmar 10mo ago
>Yes. Firefox and Chrome already offer this. yes, both use machine learning methods to translate pages. You're already using AI and don't realize it.
167.
by sigmar 10mo ago
Do you ever need a website you're visiting translated? Have you ever not understood a term or phrase on a website and had to go to wikipedia/urbandictionary/google to explain it? Have you ever wanted to do a 'fuzzy searc
168.
by sigmar 10mo ago
>This isn't about talent. It's about integration... Vertical integration isn't a nice-to-have. It's the whole ballgame. I'm going to assume there's tons of logical errors and oversights in the math, consider
169.
by sigmar 10mo ago
>We haven't built a reactor in a long time. France finished Flamanville 3 in 2024. Finland finished Olkiluoto 3 in 2022. Are those not recent enough? both were EPR designs
170.
by sigmar 10mo ago
These questions are inane. No, "all existing experts" did not retire. not making new plants was a decision made by politicians. Europe has never stopped working on creating new and better nuclear reactor designs https:/&#x2F
171.
by sigmar 10mo ago
Did it reproduce copyrighted works? Including Disney movies in training data is not an infringement of copyright.
172.
by sigmar 10mo ago
Are there any specifics about how this was trained? Especially when 5.1 is only a month old. I'm a little skeptical of benchmarks these days and wish they put this up on llmarena edit: noticed 5.2 is ranked in the webdev arena (#2 tied
173.
by sigmar 10mo ago
> The conclusion was that superforecasters' ability to filter out "noise" played a more significant role in improving accuracy than bias reduction or the efficient extraction of information. >In February 2023, Superfore
174.
by sigmar 10mo ago
Gotta auto grade every HN comment for how good it is at predicting stock market movement then check what the "most frequently correct" user is saying about the next 6 months.
175.
by sigmar 10mo ago
>In his 2019 memoir, The Threat, former Deputy Director of the FBI, Andrew McCabe, quotes Trump as saying of Venezuela "That’s the country we should be going to war with, they have all that oil and they’re right on our back door.”
176.
by sigmar 10mo ago
>If they cared about emissions they could try regulating idling, which worse emissions profiles They do, in fact, regulate idling my dude- https://portal.311.nyc.gov/article/?kanumber=KA-02222 > In New York City,
177.
by sigmar 10mo ago
Modern smartphones are very private, featuring full disk encryption, sandboxed app data storage, and mandatory lock screens. Why do you think the data stored on your phone isn't private?
178.
by sigmar 11mo ago
People hit like and then comment "ai"... I think they love being mad at ai or aren't mad enough to stop hitting like. (Just today I saw a viral video on IG of a monkey on the side of a mountain path jumping onto a man's
179.
by sigmar 11mo ago
Wow. What a post mortem. Rather than Monday morning quarterbacking how many ways this could have been prevented, I'd love to hear people sound-off on things that unexpectedly broke. I, for one, did not realize logging in to porkbun to
180.
by sigmar 11mo ago
That's a different model not in the chart. They're not going to include hundreds of fine tunes in a chart like this.

More ›