ansk

8 ms·

31.
by ansk 3y ago
The contrast between the media's depiction of today's events with this outcome demonstrates how much Sam and his aligned interests can influence a narrative. The vast majority of reporting indicated that Sam held all the leverage
32.
by ansk 3y ago
One conceivable situation in which the board's behavior could be justified is if Ilya presented them with an urgent ultimatum: either Sam goes or I go. If the rift between Sam and Ilya was growing to the point that Ilya would rather l
33.
by ansk 3y ago
> a list of numbers can not be copyrighted Any digital object can be represented as a list of numbers (this is precisely the origin of the term digit al). Since there is clearly precedent for copyrighted digital objects (media, software
34.
by ansk 3y ago
Mediocre in regard to the standard set by other research teams pulling $1B+ valuations. The OpenAI founding team and the recently announced team at xAI are what I would consider to be exceptionally strong groups. Adept, Anthropic, Inflect
35.
by ansk 3y ago
This dude very well may be better off with the 100 bucks at this point. The company is bleeding money at a ridiculous rate, so they're going to need to fundraise shortly. Given how little they have to show for their initial funding,
36.
by ansk 3y ago
These images were all generated by an identical model. The fact that this individual has convinced themself that the model is improving indicates that they don't understand how these models are trained and deployed. Furthermore, any
37.
by ansk 3y ago
This is a great rorschach test. Show these four images to someone hyping AI, and if they see evidence of a growing/emerging intelligence, you can diagnose them as being wholly unqualified to comment on anything related to AI.
38.
by ansk 4y ago
Buried in a footnote: > we now believe we were wrong in our original thinking about openness, and have pivoted from thinking we should release everything (though we open source some things, and expect to open source more exciting things
39.
by ansk 4y ago
Will Doubling Disk Size Solve Storage?
40.
by ansk 4y ago
This varies quite a bit based on the type of model. The graph-based approach has two benefits: (1) removing overhead from executing python between operations and (2) enabling compilers to make optimizations based on the graph structure. Th
41.
by ansk 4y ago
So this looks like a further convergence of the tensorflow and pytorch APIs (the lower-level APIs at least). Tensorflow was designed with compilable graphs as the primary execution model and as part of their 2.0 release, they redesigned th
42.
by ansk 4y ago
Image generative models have made significant progress over the last few years, but an often-overlooked confounding variable is that during this period, the predominant benchmarking task used by researchers (specifically those working on l
43.
by ansk 4y ago
This is exactly why Stability's role as a middleman in these model releases is an indefensible business position. The research underlying stable diffusion does not belong to them, so any company which wants to fund the training of the
44.
by ansk 4y ago
I've spent the last half hour trying to figure out what this company is, and I still have no idea. My first stop was their website and I could not find a single sentence providing concrete details regarding their value proposition. I
45.
by ansk 5y ago
Is the prevailing opinion that progress in reinforcement learning is dependent on algorithmic advances, as opposed to simply scaling existing algorithms? If that is the case, I could see this decision as an acknowledgement that they are no
46.
by ansk 5y ago
I've been working in this area for a little over a year, so perhaps the artifacts in the generated speech are more apparent to me than to others. With that said, I don't find the displayed voices to be very compelling. All of th
47.
by ansk 5y ago
You've correctly shown that maximizing the likelihood is equivalent to minimizing cross entropy in the discrete case, but frankly that is unrelated to your claim that the equivalency doesn't hold in the general case. As noted in t
48.
by ansk 5y ago
Do people not google terms before trying to speak authoritatively on a topic they aren't familiar with? The original commenter is correct, cross entropy is a generic measure of two probability distributions - in the case of maximum li
49.
by ansk 5y ago
You're clearly disillusioned with the general accessibility of ML research, but I don't think your cynicism is warranted here. Take a look at their prior works[1], and I think you'll agree they go above and beyond in making
50.
by ansk 5y ago
This group of researchers consistently demonstrates a degree of empirical rigor that is unmatched across any other ML lab in industry or academia - remarkable empirical results as always, reproducible experiments, open-source and well-engin
51.
by ansk 5y ago
Thank you.
52.
by ansk 5y ago
LinkedIn authwall, we meet again. Could someone list the researchers (if there are any, and assuming there are only a few). Frankly, it's not a great sign that the Anthropic site isn't touting the research team itself and Linked
53.
by ansk 5y ago
I can't find any mention of who currently comprises the core research team. It mentions Dario Amodei as CEO, and their listed prior work suggests some others from OpenAI may be tagging along. However, the success of this group is goi
54.
by ansk 5y ago
OpenAI is the corporate equivalent of a social media influencer. Rather than sponsored product promotion alongside a seductive lifestyle, they offer Azure product placement alongside trendy AI research and, now, futuristic startups.
55.
by ansk 5y ago
The preeminence of either GANs or diffusion models is an ongoing debate, with evidence trickling in paper by paper. To claim diffusion models beat GANs is like claiming the Red Sox beat the Yankees because they happen to be ahead in the thi
56.
by ansk 5y ago
For an actual example of this more cautious framing, see another paper (also by OpenAI, also comparing two classes of generative models), https://arxiv.org/abs/2011.10650 .
57.
by ansk 5y ago
Why broadly claim an entire class of generative model outperforms another? They simply show that their particular implementation of a diffusion model attains better performance (for some metric) than the current best implementation of a GA
58.
by ansk 5y ago
Further off-topic, but do people actually consider this to be beautiful design? Looks like a rendered markdown document with MathJax and green headers. Perfectly appropriate for the content of the post, but beautiful isn't the first
59.
by ansk 5y ago
K (as well as X and V) is supposed to be a fractal, so the pattern isn't just arbitrary. X and V are too dense in my opinion and could use a lower branching factor, maybe base 2 for consistency. K is better in this regard, but they d
60.
by ansk 6y ago
Spotlight correctly interprets numerical expressions, but it also treats them as search queries for the entire filesystem, resulting in an expensive retrieval process after every keystroke. It's the most convenient, so I use it anyway

More ›