iandanforth

"Although not one of the five Nobel Prizes established by Alfred Nobel's will in 1895, it is commonly referred to as the Nobel Prize in Economics, and is administered and referred to along with the Nobel Prizes by the Nobel Foundation. Winners of the Prize in Economic Sciences are chosen in a similar manner as and announced alongside the Nobel Prize recipients, and receive the Prize in Economic Sciences at the Nobel Prize Award Ceremony."

- Wikipedia

This is a good point! I've got a few decades of tolerating neosporin for normal uses so wasn't overly concerned.

Yes, mask first, this is a second line of defense. (Again, maybe, could do nothing or be harmful)

I will admit to following the swab-nose-with-neosporin protocol following a previous mouse study with similar results. I use this during travel and have had no short terms ill effects and caught no infections while following it. (Not a doctor, not well controlled, just a random internet guy).

https://www.ncbi.nlm.nih.gov/research/coronavirus/publicatio...

Going plastic free is very hard today. I'd be interested to know if people have managed inroads on this problem in their own lives.

Important distinction is that tincans is not speech to speech. It uses a separate turn/pause detection model and a text to speech final processing step.

Meta: dang here is incorrect.

Each post could be the first time someone is encountering a Wolfram post. If they haven't built up defenses to this kind of self aggrandizement they might believe his words. This, and every piece from him deserves a community note.

Just because you've seen a lot of misinformation from a source in the past is no excuse to stop labeling it as such.

How are people expected to answer the usage questions accurately?

"I'm all for smart things ..." Root cause found. Cannot fix, pebkac.

Answer: No.

You have to be very stingy with your definition of history and writing to ignore the fossil record as a recording of historical events. We'll revisit the question in several hundred million years and see how oral traditions are doing.

I once had to build a complex NN based system for this exact task. We also pitted this system against a consultant building a basic xgboost classifier working with a series of sliding windows over sensor data. Xgboost won. It was also faster to run and faster to train.

I'd still love to see how a VLM did with sensor data as it's often quite obvious to spot anomalies visually for humans. Especially if you're allowed to do comparison overlays.

Exciting and very cool! I look forward to the continued improvement in this area. Especially when the loop is closed within Sakana and you can say "this discovery was made by The AI Scientist" as part of another paper.

If I might offer some small feedback on the blog post:

- Alt-text and/or caption of the initial image would be helpful for screen readers

- Using both "dramatically" and "radically" in one sentence to describe near future improvements seems a bit much.

- When talking about the models used, "Sonnet" could either be 3.0 Sonnet or 3.5 Sonnet and those have pretty different capabilities.

Thanks again for the impressive work!

No mention of suing Apple, no mention of reducing their cut. Doesn't really seem like Patreon has got creators backs.

The video and the pitch strike me as odd. The value proposition here is to grow the market for small planes by making them easy. You still need to be rich so this is basically an expensive toy. I'm not sure comparisons to what came before matter, the people who will buy this don't know or care about the existing way to control a plane.

All you have to do is show how easy your system is. To which end, why the hell would you put two separate hand controls rather than a steering wheel that also goes in and out and some pedals for speed up and slow down. It it's as easy as a car, make it like a car.

I would be more comfortable with this technology if it had saw-stop like behavior. If I sneeze it has to be fast enough to safe the dental tool before I impale myself on it.

For some reason I thought this was about simulating the fabric in a physical sense. Having watched the videos from the supplemental material (downloadable from the SIGGRAPH paper page) it appears to be about "just" rendering. The fabric does look nice though for some reason they chose very shiny fabrics for their video.

It's free but the Amazon reviews are ... not favorable. I'd love to know of books or resources like this that are good collections of real AI/ML/DL interview questions though!

"Please use a valid institutional or company email address."

This is obnoxious. Please remove this unnecessary roadblock.

Why does it weigh 100lbs?

nit: I find the writing in this post very distracting. (Grammar and style pet peeves)

Luckily, it is now trivial to drop the post into Claude and say "Re-write this without <list of things that bother me>"

So, just in case you also felt like you were driving over a road filled with potholes trying to read this post, don't just click away, have your handy LLM take a pass at it. There's good stuff to be found.

While this article presents a system it doesn't present any results. Does this modification to SRS help? Which type of student does it help? How large is the effect if there is one?

I haven't read the larger pdf of which this is a part so perhaps someone who has can provide a pointer to some results.

And yet the consequence of letting people like this run your security org is that it takes a JIRA ticket and multiple days, weeks, never to be able to install 'unapproved' software on you laptop.

Then if you've got the software you need to do your job you're stuck in endless cycles of "pause and think" trying to create the mythical "secure by design" software which does not exist. And then you get hacked anyway because someone got an email (with no attachments) telling them to call the CISO right away, who then helpfully walks them through a security "upgrade" on their machine.

Caveats: Yes there is a balance and log anomaly detection followed by actual human inspection is a good idea!

Open source is great and new things are great and pursuing your passion is great. The rhetoric here however is lacking. Specifically the argument is "google money bad" but the authors don't provide specific examples where google money has caused a technical decision they disagree with.

The semantic equivalence of possible outputs is already encoded in the model. While it is not necessarily recoverable from the logits of a particular sampling rollout it exists throughout prior layers.

So this is basically saying we shouldn't try to estimate entropy over logits, but should be able to learn a function from activations earlier in the network to a degree of uncertainty that would signal (aka be classifiable as) confabulation.

Are you a lawyer? Can you point to the relevant statutes? I don't ask this flippantly as there exist many forms (private / public / academic) of library which allow for both physical and digital lending of owned assets and are not subject to lawsuits like this. It's not at all obvious that IA's interpretation of the law is in error.

If you enjoy this topic I also highly recommend "A Brief History of Intelligence" which goes into quite a bit of detail, is very readable, and ties in directly to the near term future of what intelligence will mean in our world. Really a very good book!

I don't envy Australian's most things but this is one worthy of it.

One point not covered but clearly understood tacitly by the author is that the representativeness of a fact degrades over time. While it may be true that someone did something 10 years ago, that fact may not be a true reflection of their character today. Thus a statement may be technically true, but practically false, and thus should be considered defamatory (under the modern American definition).

I think the only way I would trust this is if they explicitly described how they would combat 5-eyes surveillance. If you're not willing to acknowledge that the most dangerous foe of privacy in the western world is the governments of the western world then why should I believe anything you have to say about your implementation?

I'm amazed they were able to pull from Planet. As impressive as OpenAI is, Planet is such a stellar company with the right mix of good intentions/impact and profitable product market fit that it's hard to imagine a better place to work right now. That said, maybe Planet's success made it less attractive to someone looking to take on new and monumental challenges.