IME having a "rubber duck" for debugging & planning can be helpful. Having it talk back is a significant boost. Having it write the code isn't. AI is nice at helping me explain the problem I'm trying to solve to myself. It's much harder to write clear, unambiguous, succinct English than any programming language, but being forced to do so to explain something to an AI helps clarify my own thoughts. The AI's output isn't nearly as relevant, though it can be helpful since it can run its own searches across the code & invalidate my assumptions sometimes.
An extra tier of standard library which can make breaking changes, perhaps. Rust's stability guarantee for std means cryptography really shouldn't go in there, since sometimes algorithms & protocols get broken (DES, MD5, SHA1, etc.) and need to be removable. Without breaking changes you get stuck with security vulnerabilities, if not from cryptography then from other poorly-designed APIs.
> I'm just not sure how this gives me control of my information, whether I want it sent or not to Google, and if they're retaining it for training or not.
That's not the goal. Turning your information into their information is the goal. If training an LLM on some data isn't copyright infringement but instead makes it a brand new non-derivative work, then training an LLM on your personal information arguably means it's no longer your personal information, but instead a legally distinct work.
Batteries & solar (with sunlight) can black start if they have grid-forming inverters. Grid-following inverters are cheaper, but there's nothing physical or engineering-based that prevents using grid-forming inverters on enough renewable sources to black start when needed, just the economics of cheaper installations.
Those were Trump's proposed solutions to the COVID-19 pandemic.
Their post is quite clear: the US has an utterly incompetent administration which is actively opposed to the best known method for preventing the spread of deadly viruses: vaccines. Funding for monitoring the spread of diseases has also been cut. The World Cup is a large event which will bring people from all over the world into the US and into close contact for multiple days, right as a new dangerous virus with a long incubation period has been detected spreading from person to person.
The post is unnecessarily political, and is written quite unclearly. The administration certainly will fund a new vaccine if the situation merits it.
The Covid mRNA vaccine, while lifesaving, is not without safety issues. It increases carditis risk substantially[1] (5x the baseline risk in males), and I personally experienced this effect for a month. It is best not to neglect such safety issues, ideally so they might be addressed in future protocols.
No vaccine is without safety issues. That's why so much research has been done to evaluate the safety issues. The consensus conclusion is the benefits of vaccination far outweigh the risks in this age group, including because the risk of myocarditis or pericarditis is higher for those who get COVID. For examples more recent than 2022:
"Although our meta-analysis revealed a higher risk than previously reported of myocarditis/pericarditis attributable to mRNA COVID-19 vaccines (among adolescent and young adult men), this does not negate the recommendation of COVID-19 vaccination for this population. Our previous study showed that the benefit of receiving the BNT162b2 vaccine or the mRNA-1273 vaccine was much higher than the risk of the vaccination across all age group and sex.59 Though the AR of myocarditis was assumed to be 12.1 per 100 000 for the primary series of the mRNA-1273 vaccine in men aged 18-29 years for the original benefit-risk calculations, if we updated the AR to 22.26 (2.84 after first doses plus 20.0 after second doses) per 100 000 for the primary series for men aged 18-24 years, based on the results of this current systematic review with meta-analysis, the benefits of vaccination would still far outweigh the risks in this age group." - Epidemiologic Reviews, 2025, 47, (1), 1–11 https://doi.org/10.1093/epirev/mxae007 Advance access publication date: December 13, 2024
"Compared with unvaccinated groups or unvaccinated time periods, the highest attributable risk of myocarditis or pericarditis was observed after the second dose in boys aged 12-17 years (10.18 per 100 000 doses [95% CI, 0.50-19.87]) of the BNT162b2 vaccine and in young men aged 18-24 years (attributable risk, 20.02 per 100 000 doses [95% CI, 10.47-29.57]) for the mRNA-1273 vaccine ... Young people’s risk of developing myocarditis is higher and longer lasting after covid-19 infection than after vaccination against it, the largest study of its kind suggests ... Over a six month period the researchers estimated that covid-19 infection led to 2.24 extra cases of myocarditis or pericarditis per 100 000 children and young people. This compares with 0.85 extra cases of myocarditis or pericarditis per 100 000 children and young people in those who were vaccinated." - BMJ 2025; 391 doi: https://doi.org/10.1136/bmj.r2330 (Published 05 November 2025)
The head of the Department of Health is solidly opposed to the development of vaccines. And the COVID mRNA vaccine increases carditis risk less than getting COVID without having had the vaccine does, and decreases the risk of severe carditis if you get COVID after the vaccine. So that's not a great argument against it.
Also, I just checked, and LinkedIn's privacy policy page doesn't contain any information about who viewed my profile in the last year. No usernames, no company names, it's just a generic privacy policy. So the data isn't there either.
I've got a 61mp camera, and an RX 7900XT. It takes about 15s/picture for DXO to denoise, which is a lot longer than people are willing to wait on a phone to take a photo. Topaz is even slower. A cloud service could be used to do it in post, but someone has to pay for that.
Yet in modern computer games, modern graphics cards denoise a scene in real-time at 60 frames per second using machine learning models [1][2] while doing all the other rendering at the same time. Granted, that's ray tracing, and the resolution is lower, and they technically cheat by using additional information, but it might be that DXO is not optimized very well.
Games will typically un at 4k or less which is about 8mp on the other hand it’s difficult to buy a stills camera with less than 20 mp, and >40mp is common. Most algorithms are n^2 in graphics as well so we wouldn’t expect a linear speed up. I’ve tried dxo, Lightroom, and topaz they all perform about the same so I don’t think it’s particularly unoptimized.
I've been learning FreeCAD, while it's still more frustrating than Fusion or SolidWorks it's much better than it used to be pre-1.1, and it's FOSS. Also constraint-based, I've been using the new spreadsheet view as the source of all constraint dimensions, with parts derived by binding to top, front, or right-side orthographic "master" sketches. Much like hand-drawn design, where you draw the orthographic views and use those directly to create an isometric view.
reply