A recent WSJ Journal article [1] summarized some of the positive benefits being derived through the use of AI. Here is a bulleted list of the benefits gained.
AI can predict protein structures in hours rather than years, and AlphaFold has mapped nearly all 200 million proteins known to science, accelerating disease research for millions of researchers worldwide.
AI helped enable the development of a COVID-19 vaccine built with a computer-designed protein.
A drug designed by generative AI improved lung function in patients with a fatal lung disease in a randomized trial.
AI identified a new class of antibiotics capable of killing drug-resistant staph bacteria.
AI-assisted mammography detected 29 percent more cancers, missed fewer aggressive cancers, and reduced radiologistsâ workload by 44 percent.
An OpenAI model solved an 80-year-old mathematical problem originally posed by Paul ErdĹs.
Professionals using AI completed writing tasks 40 percent faster while producing work rated 18 percent higher in quality.
AI assistants increased customer-service productivity, with the largest gains among newer and less-skilled workers.
In an Argentine study, AI closed roughly three-fourths of the productivity gap between more- and less-educated adults.
Thus far, AI adoption has shown no detectable overall effect on employment, earnings, or hours worked, and the feared widespread AI-driven layoffs have not appeared in the data.
There are good things but without regulations the good can turn into bad especially in the medical and health field.
Rest in terms of technology field it just is starting to get its feet wet. There will be widespread security issues that one can possibly imagine. The blast radius of AI Security flaws are pretty much infinite.
If our government cannot agree in 2026 that the Measles vaccine is helpful and took decades to recognize the risks of tobacco and lead, what are the odds that government can regulate AI in any meaningful fashion?
Agreed, but the devil is always in the details. You have probably seen it, but if not, there is an extensive discussion of regulating AI in this thread.
Unfortunately good news doesnât grab attention like AI is going to kill us all. If we want to know what is going on we have to get into the weeds. We canât rely on headlines in the news or comments on social media.
Chances are some of us may know someone thatâs already be helped by AI:
âAidocâs platform is running right now in more than 1600 hospitals worldwide, analyzing over 70 million patient cases a year. It holds 32 FDA clearances, the most in its category. When a patient gets a CT scan in an emergency department that runs Aidoc, the AI analyzes the images as they complete and flags time-sensitive findings, brain hemorrhages, pulmonary emboli, aortic dissections.
So, the radiologist sees the most urgent cases first, rather than in the order they arrived at the ER. When minutes matter, that patient lineup can be the difference between a good outcome or a catastrophic one.â
From KFFâs The Business of Health with Chip Kahn: When Minutes Matter, What Is AIâs Role?, May 12, 2026
The best use for me has been using the Gemini API in DevonThink to help with recruitment and processing CVs. Itâs saved me hours.
I also really love the Superhuman Business plan. Having a reliable âthinkingâ AI trained on my email is the most useful tool I have at work now. The quick email processing is a nice bonus, especially the âquick replyâ feature, which creates a draft reply to every email before I open it.
That is an interesting use case. I have Claude connected to my DEVONthink research database. Can you elaborate a little on how you have Gemini connected and how you use it for processing CVs?
Sure! I use the API key, which works in the same way as Claude but costs much less. I tried both and got very similar results so went with the one thatâs costs less.
I have a database for recruitment, with CVs filed by specialisation (e.g. math, ML, databases etc.) Then, when I have to make a new appointment, I ask Gemini Pro to look at a job spec/syllabus and then to analyse the CVs in one of the categories (I select the files to analyse, including the syllabus/job spec). I ask it to tell me the best fit for the job and the results have been amazing. In the last two instances, the job matched the top selected profile so well that the candidate I interviewed got the position in both cases. Normally, I interview at least 5-6 people before I find a fit.
I also use it to sort, rename and organise the CVs when I download them from our recruitment platform.
Thank you, that is extremely helpful information. Iâm going to pass this on to my HR department. Thank you for taking the time to explain how youâre using itâ much appreciated.
There is good in AI, itâs helped me tremendously with several projects.
I was using a platform for event registration that was extremely complicated (7 pages) and it had to account for single, married, children, along with which tracks, each person will attend, etc. My formula kept glitching and I couldnât figure it out, it found out for me in under a minute, and helped fine-tune some adjustments. On top of that, probably a critical time-consuming piece for me was the data export. It exported and matched up everyone according to their track for the weekend. The platform doesnât typically support âfamily styleâ registration so in years past, I had to be creative which meant it was a great user experience but a really bad export /finalize report experience.
Cleaned up my firewalls / network policies. It found holes that I thought I had closed up.
The list goes on. AI wonât kill usâŚlack of discernment will.