drtournier

I wonder how much it is related to the invisible watermarking

PC games were the last reason why I still used windows. Now that Proton works wonders in Linux, I hope I don’t need to ever come back again. Great times with 98, xp and win7 that were snappy, easy to install and with no slop

Interesting question - what I learned doing that is that sol ends up injecting its thinking traces as code and shell terminal comments. YMMV, but in my case what I saw was sufficient to use with my routine work in my projects

pi sends 1k (or less) -> https://github.com/earendil-works/pi/blob/main/packages/codi...

My $20 sub using gpt 5.6 sol thinking-off lasts for hours using pi.

Always loved to have an e-ink reader but the walled gardens always made it so hard. Until I bought the xteink x4, as skeptical as I could be. What a great little device. https://www.xteink.com/products/xteink-x4

$> telnet tsunami.thebigwave.net

The one and only. Still online to my surprise.

JavaScript is not a great language for web development either, yet…

OP thank you for this generous text. I would add that these tips are amazing for people with post-covid/long-covid/brain-fog too

I just discovered that 90% of all credit card activity still rely on mainframes https://www.mordorintelligence.com/industry-reports/mainfram...

<quietly turns airplane mode on for good>

Thank goodness it is a bug. For a second, I thought I was on the new squid game.

And it worked. Congrats for reaching Front Page, Bruno!

This article contains potentially harmful advice. I would never expose my eyes to direct sunlight - The risk of permanent retinal damage is insanely high, even if you gaze at the sun for a few seconds straight.

I was an user of zeppelin for a couple of years (v0.7-0.8), we used it to run Scala Spark and the UI has a lot of bells and whistles to make it easier to use Spark, display spark dataframes and simple dataviz features out-of-the-box. It is a bit relatable to the notebook experience you would have in databricks.

$>telnet tsunami.thebigwave.net

Coffee and tobacco consumption is a known correlation (one of many refs https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4158991/). Curious to see if authors considered that, as tobacco is one of the main aggressors of vascular tissue, which causes dementia in the long run.

EDIT 1: Just in time, UK Biobank (almost 500.000 people) study pointing to increased longevity with coffee consumers, an interesting controversy that just doesn't fit with their smaller study conclusion (although overall longevity is a different outcome, you would expect to find more dementia in older people) https://www.hsph.harvard.edu/news/hsph-in-the-news/coffee-lo...

Our case took around 14 months for the EB-1 from the first call to lawyer to the green card in our mailbox.

Glad you made it OP, congrats! As an immigrant in the US I went through the "Extraordinary Ability" path and the process was a journey of pain and anxiety. Many roundabouts with a lawyer that promised me expertise in the process and it ended up with a NOID (Notice of Intention to Deny - something like that) from USCIS. My wife and I had to study the USCIS manual page by page and re-wrote almost all documentation from the lawyer. It wasn't easy but we made it.

Just read two paragraphs and sounded like someone asked ChatGPT to write this thing as a Joe Rogan podcast.

This is awesome. Long time I wanted to pre-train a language model in Brazilian Portuguese and just got a 24GB 3090. This is the perfect timing. Thank you for sharing your notebook.

MKML == abstractions and wrappers for GGML?

The idea is cool. But - which existing language is not neurosymbolic?

"come back to the office we got kombucha on tap"

https://twitter.com/Jobvo/status/1496507164445356032

Some things that help me a lot:

- Go walk outside

- Consume media / go to events with topics that are VERY different than my field of expertise

- Do any kind of art

- Play like a kid

- Last but not least, my favorite: Have a lot of bad ideas. https://seths.blog/2018/06/the-two-simple-secrets-to-good-id...

That's the whole point on this NFT conundrum, right? One can post something without even have to confirm that they are the owner by rights of what they are selling. https://www.vice.com/en/article/wxdzb5/more-than-80-of-nfts-...

Thanks for the relevant comment - it must be intriguing for anyone who is not used with hearing about the “real world evidence” term in medical sciences.

Usually the highest level of evidence is attributed to randomized controlled trial studies. As the name says, they are performed in a high level of control, to evaluate in a more “insulated” way, the effects of a drug/intervention in a given outcome (reduce death, cure a disease, etc). Unfortunately, we can’t always relate these academic studies with the “real world”, due to an almost infinite number of confounding variables, or in the worst case scenario, if a scientist is manipulating the experiment environment (e.g. selecting a patient population with more likelihood for the desired effect).

This is one of the reasons why you will see with increasing frequency the terms “Real World Evidence (RWE) study” in reputable journals, such as the New England Journal of Medicine, or The Lancet.

These studies occur taking data from epidemiological surveillance, or patient medical records, and usually do a “reality check” on data found in these trials.

A lot of these RWE studies are famous to find problems that trials missed. Here is a classic example https://thalidomide.ca/en/what-is-thalidomide/

I am an MD, Data Scientist and epidemiologist. There are three pieces of information which makes me feel something fishy about the study: 1. Sample sizes for all selected studies are very small, and even though they used mixed effects models to treat confounding, there is no way to check what the authors did, since they didn’t shared the code or the data, which is inadmissible since this is a meta analysis and there is no private data involved 2. Any tree chart of Risk Ratios or Odds Ratios that show drastic positive effects must raise one’s eyebrow about partiality in the authors 3. Real world data from Brazil, where they have great death data surveillance AND mass-distribution of ivermectin didn’t show any excitement. Brazil just crossed more than half million deaths as of Jun 19th 2021, and there is a high likelihood of surpass U.S. in deaths.

Based on what we already know about vaccine and PFF2/N95 mask efficacy and safety, it becomes hard to defend approaches such as ivermectin, instead of promoting practices and technologies that showed undeniable success in the real world, instead of doubtful medicines supported by obscure statistics and ideological discourse.

Solution works really well to databases which will not be updated frequently, like a standalone site.

Although one should be aware of one very important git behavior - git does not diff binary files (like SQLite dbs). That means 2 things:

1. Each db update will generate a new file in git, maintaining the whole old file in history, instead of the diff in bytes. This will accumulate a lot of clutter in the repo

2. As git does not diff binaries, there is a very small risk of corruption (especially if you work in multiple OSs, because of CRLF)

Ref - https://robinwinslow.uk/dont-ever-commit-binary-files-to-git