radus

I'm sure this type of integration will happen, but... isn't this exactly how AGI would "escape"?

Yes

Did you work in the meat packing department of Wendy’s? Your experience may hav e been a tad different..

https://excessups.com/ sells batteries and UPS units. The downside is they can be quite slow to ship.

The way they're going about this investigation is reminiscent of how we figure things out in biology and neuroscience. Perhaps a biologist won't do the best job at fixing a radio [1], but they might do alright debugging neural networks.

1: https://www.cell.com/cancer-cell/pdf/S1535-6108(02)00133-2.p...

So you’re saying there’s a chance?

This is most threads on topics outside of the CS sphere

Yeah, sci-hub is great, but I'm pretty sure academic research would be fine without it... on account of having been a biology PhD student both before and after the advent of sci-hub.

The research described in the pre-print aimed to understand the mechanism of action behind the decrease in pathogenicity observed with Omicron strains, and specifically, whether mutations in the Omicron spike protein can account for that decrease. The experiment they carried out is about the most straightforward one you could draw up to begin to answer that question, and I would never even think to characterize it as a "gain of function" experiment. If you want to classify this study as GOF then you might as well do the same with any study of a virus where even a point mutation is made - good luck doing any virology under those conditions.

I really really wanted to like mypy but my experience with mypy and Django has been very poor - it is slow, type inference is not good and most of the errors are false positives. Perhaps I’m spoiled by Typescript or django-stubs is just not quite mature enough.

That level of jaywalking enforcement sounds dystopian, no thank you.

> I program the happy path then write tests and use them to help uncover edge cases.

This approach resonates with me as well. I would add that writing tests when investigating bugs or deviations from expected behavior is also useful.

More even, billions+

How do you CI/CD assertions in papers using animal models in experiments that take months to years?

Unfortunately, abacus side-channel exploits are numerous as well

You can opt out of PEP-582, and in PDM 2.0 (just released), it becomes opt in.

Yep agreed. For me it's welcome because this latest generation of churn feels like it has leveled up DX considerably. The churn in between Backbone up until ~ Angular 1 brought a lot less value in my experience at least.

Yes you’re definitely right - I’ve recently started using SeleniumBase for Python which includes a few extra niceties. I was more asking about the part where you compare between releases (or test suite runs) to see if anything changed. I suppose you can just use compare hashes, but I could imagine also having some feature that highlights changes on the screenshot.

SeleniumBase has a feature that looks at markup and which stores screenshots for visual comparison to a set baseline: https://seleniumbase.io/examples/visual_testing/ReadMe/

Found this blog describing use of SeleniumBase for that and OpenCV for screenshot comparison as another layer: https://blog.streamlit.io/testing-streamlit-apps-using-selen...

What are the OSS solutions for this?

I think it would be ideal if the increase in responsibility was accompanied by some actual reward, so the “congratulations” would be more genuine. I don’t know what kind of reward would make sense - money makes sense, but who’s money? Maybe you get a voice in the development and direction of the language and ecosystem? That would also constitute more responsibility but it might feel less burdensome and give you a seat at the table for the next such decision.

In that scenario though I think there are reasons why one might avoid package repositories such as pypi. For instance, say a censorship avoidance package is popular on pypi - what happens when whatever authority comes knocking?

I would think that it would be more secure to provide anonymized distribution as well. Of course, that means that you lose some convenience and reach, but that’s a common trade off in scenarios that have elevated security needs.

The vast majority of scientists I know (I'm a scientist also) are self-doubting to a fault and often feel stupid even if they're not.

“Please extract this to a separate function”

Good summary of other changes: https://www.infoworld.com/article/3660550/the-best-new-featu...

And of course, the docs for the fuller story: https://docs.python.org/3.11/whatsnew/3.11.html

Definitely looking forward to the release later in the year

They’re using the requirements of the python standard library logging module as an example to explore different patterns to accomplish the desired feature set. This module does need to be flexible for obvious reasons.

Allow me to contextualize and re-state:

> Unless you can draw a line all the way from diagnostic use to results to improved outcomes, it's just data.

This is why we do trials, to figure out if we can draw a straight line from results (pre-clinical) to outcomes (clinical). This is the question that the cancer diagnostic trial will hopefully answer - can we improve outcomes using these diagnostics?

I don't completely see the relevance of your story about inhaled insulin. As far as I can tell it was not better than standard of care in a meaningful way, whereas with this trial there really isn't a standard of care to speak of, and this is going to be the largest such trial to date. Plenty of drugs fail to make an impact in the market, for a variety of reasons, which can't always be predicted ahead of time.

Elsewhere you write that this is not going to provide actionable information. I largely agree with this statement - as of today. I also think that this trial will provide valuable information about the landscape of cancer diagnostics in general, information which may be increasingly relevant in the future and may turn out to be actionable. This is the most relevant quote from the source article that speaks to why they are doing this:

> Q: Given the doubts, why proceed? > A: The exciting thing is, we don’t have screening tests for the lethal cancers like pancreas and ovarian. We’re desperate. But we have to put our emotions aside and do our due diligence to evaluate these technologies and be able to speak with confidence about what this can and can’t do.

For 75 million dollars, the cost of a year of maintenance for a handful of fighter jets, I think it's worth it, and I'd much rather have the NCI running this than a private company. If this trial completely falls on its face and public money keeps being expended in this direction, then I would certainly revisit my opinion about the worthiness of trials of this nature.

This is why we do trials.

Calculate a rough semantic similarity score across all your snippets, and pay out a fractional reward to all originating codebases.

I think the bigger problem is that it will almost certainly lead to a proliferation of giant snippet spam repositories.

Thanks!

Question for you or anyone else: what is your sense for whether or not Swarm will continue to be developed? I recall some noise about it being canned.

Third, for me at least: Remote Extensions. That suite of extensions including SSH WSL, and Container support made things I was already doing frictionless.