I use gemma3:1b model (well, gemma3n:e2b since today) to summarize articles in my RSS reader. Works extremely well for such a simple task and runs on CPU on my hetzner server, so I don't have to pay electricity bill for running it on GPU at home
Setting up Cilium with a pool of ips to use as a lb: 45 mins.
All in: ~5 hours and I'm ready to deploy and spending 300 bucks a month, just renting bare metal servers.
AWS, for far less compute and same capabilities: approximately 800-1000 bucks a month, and takes about 3 hours -- we aren't even counting egress costs yet.
So, for two extra hours on your initial setup, you can save a ridiculous amount of money. Maintenance is actually less work than AWS too.
"The Linux Programming Interface" is excellent, and very thorough (1500 pages!).
But it's already 12 years old. Are there important innovations in Linux and its APIs from the last 12 years that would be interesting to learn, and how ?
See also the leak[1] of big oil PR maestro Rick Berman's 2014 speech to oil execs, pleading for funds to combat rational debate and reliable information about fracking.
“Think of this as an endless war,” he told them. “And you have to budget for it.”
> He said the industry needs to dig up embarrassing tidbits about environmentalists and liberal celebrities, exploit the public’s short attention span for scientific debate, and play on people’s emotions.
> “Fear and anger have to be a part of this campaign,” Berman said. “We’re not going to get people to like the oil and gas industry over the next few months.”
> Berman also advised that executives continue to spend big. “I think $2 to $3 million would be a game changer,” he said. “We’ve had six-figure contributions to date from a few companies in this room to help us get to where we are.”
> But always cover your tracks, he suggested, adding that no-one is better equipped at doing so than his firm. “We run all this stuff through nonprofit organizations that are insulated from having to disclose donors. There is total anonymity,” he said. “People don’t know who supports us. We’ve been doing this for 20-something years in this regard.” [2]
Plenty more in the transcript at [1], and the article at [2]!
i'm late to the comments but hopefully this helps someone:
i struggled with DP as much as anyone. i read all of the standard resources (CLRS, vazirani, kleinberg, etc), watch all the youtube videos, did all of the practice problems in the books, and still couldn't solve the kinds that are asked on interviews. i even went as far as emailing kleinberg for help.
what made it basically unconsciously fluent for me (i.e. i can read a problem statement and sketch out the recursion and subproblems in about 60s and then just perform fixup) was doing hordes of them on leetcode in preparation for a FB interview. it got to the point where i could solve hard ones in about 5 minutes using either bottom-up or top-down (i.e. memoization). so if you're struggling with DP for interviews my suggestion (which is basically the standard suggestion) is to just grind the problems on leetcode.
and contrary to popular belief they do come up outside of interviews - i had to solve a circuit synthesis problem last week and it turned out to be basically DP substring counting problem. took me all of 5 minutes.
I have archives of comp.lang.c group going back 35 years (1986-2021) and when I saw this news I decided to prioritize this one and just started synchronizing it into my archive.
In a couple of hours, you should be able to see all of the posts ever made to: comp.lang.c group over at my Usenet Archives:
BTW. I just reworked entire architecture of the site, which unfortunately also means I have to re-synchronize over 1 billion posts back into the archive. That will take some time, but anyways, just wanted to tell you that I prioritized this group. Enjoy!
External SSD is my default setup. I encourage it in the office as well. Even my gaming machine is setup at home this way. When popping over a mates I often just plug it out and stick it in to their machine.
It's very useful to be able to seperate data from the hardware. If something is wrong with a machine I can just plug straight in to another one. An SSD is much easier to carry than a laptop.
The other problem with internal hard drives is when they get stuck inside e.g. with MacBooks being sealed these days. External SSDs can offer great flexibility and a trade off worth considering. I can imagine people having security concerns etc.
Disappointing to hear this about the new M1s.
I recommend people interested in dipping their toes into Linux Gaming trying this approach:
If nothing else this should do quite a lot to reduce carbon emissions, and I'm only half joking here.
Now if we could change fashion from fast fashion to focus on durability that would be an even bigger change. And it is not completely unrealistic:
While I have no hard and fast rule to go by on how exactly this could turn out I do sense a huge opportunity for one or more smaller brands to more or less utterly disrupt big chunks of the clothes manufacturing industry by providing reliable, repairable and nice garments.
Such efforts exist, but I'm waiting for something to hit mainstream:
As it it happens I work next to a Norrøna store, and they are serious about: When I was there there was a long row of garments waiting there, and some of them must be well above 20 years old based on how the logo and the style has changed over the years.
The Stackoverflow schema is decent. Especially for someone that is a relative amateur to study. It's nothing special or complex, however it's a good example of something that actually works well in practice (and at massive scale). The CRUD-CMS Q&A style lends itself nicely to a basic db schema that is easy to get your head around at a glance.
I've demoed my $25 Anker or mpow earbuds (the Airpod-likes) to many coworkers sporting Airpods and they're universally astounded by the quality of the cheaper buds. I've converted quite a few people who lost their Airpods.
The chead buds don't sound quite as good as Airpods, but they're 85% of the way there, fit better, can actually handle the rain and sweat without breaking down, can be used during exercise without falling out, and they cost a fraction of the price.
I run an email security/monitoring business [0], and we hear this issue with Gmail a lot, in fact: we have been bitten by it as well during the early days.
The problem at hand is that it is really hard to debug situations like this. Google won't tell you why a particular email is blocked, because that information will be immediately exploited by spammers.
We build our software suite because we found that mistakes in email configuration are easy to make and really hard to identify.
I did a quick scan of rafa.eu.org and found the following:
- SPF is setup [1] with a neutral 'all' mechanism, which basically disables SPF for this domain. Hence, DMARC will set SPF as 'pass' even though SPF did nothing here to help the spam algorithm to assess the sender.
- A DMARC record exists and has a valid syntax [2], but the proposition (p value) is set to 'none', which basically disables DMARC altogether. It will enable reporting, but nothing more than that.
- I don't know you DKIM selector, so I can't assess that, but make sure the DKIM signature address is aligned (using the same domain name).
- The TLS configuration of your email server uses a self-signed certificate and is very much outdated (it offers SSLv3 and RC4 ciphers). I used testlssh.sh [3] to check this.
So yes, you technically did setup TLS, SPF and DMARC, but in all 3 you have configuration errors. Running a mail server in 2019 requires a bit more work and maintenance than it used to be 10 years ago, but the reason is spammers, not Google. Remember: it is in Google's best interest to have as few false positives, as it ultimately benefits the users. It's just really frustrating that it is hard to figure out why Google is marking your email as spam.
I just set up my own Bitwarden server the other day using bitwarden_rs, a third-party implementation written in Rust.
It basically gives you all of the premium features for free, as opposed to the official server which requires a license.
I really wanted to run the official server, but they offered no option of a lifetime license (only a yearly license). For what it's worth, I would have been willing to pay a lot more for a license that never expired.
The whole reason I'm hosting the server myself in the first place is because I want _full_ control, so a subscription based license doesn't really fit well there.
Given that the project is licensed under the GPL, the license is effectively a donation anyway, so I hope they consider offering a lifetime license for those who want to self-host.
Kind of missing any discussion of options like pfSense, or the other fancy BSD powered options. The discussed routers are the sort of options I would look at for buying for a non-techy family member, where I log in and update the firmware for them once or twice a year. But for people looking for a little more power and / or configurability, just pick up a little low power AMD x86 box and put some real firewall software on it. Add a little Ubiquiti access point and you've got a great, solid setup.
I recommend activity-watch[0] if anyone wants a good, open-source, at-a-glance system of tracking for their own computer usage. It's definitely helped me rebuild a memory of my activities on days that got away from me.
It's not just FB, my Samsung S9 came preloaded with 'undeletable' Microsoft apps too. But this is nothing new, Samsung phones have came preloaded with bloat since forever.
When buying a new phone I always spend some time deleting all Samsung, Microsoft, Facebook and carrier related apps. Yes, you can delete 'undeletable' apps through ADB, without rooting the device.
And note the comment that Transcend released the firmware (though that link is offline today.)
However I guess there really just aren't all that many use cases for this kind of a server. Or maybe it has been superseded? One page[1] lists it as weighing just 2 grams, compared with the weight of 2.5 grams for a U.S. penny. That's gotta have some sort of applications. Why isn't this more popular at $30, or what are people using instead? Is it like, impossible to get any ports out of it, even a single serial debugging port via soldering, so you can't connect it to anything else?
The book focuses on classical (statistical) methods of forecasting. In this sense, it provides the fundamental notions needed to deal with practical problems. Real world problems are much more complicated and first of all because of the natural of source data which is not limited by univariate numeric data. In most practical cases the success depends on the ability to extract (manually or automatically) good features from heterogeneous data sources. There exist the following frameworks for that purpose:
> give me Visual Basic 6's UI builder and all the greatness of a 21st century machine, so that I could finally whip up an UI without yelling profanities at the CSS reference [...] that clearly just wasn't built for this
Can I interest you in a VB6-y design/code/deploy tool with no CSS or JS? It's all in Python, even in-browser: https://anvil.works
Pretend http://facebook.com/l.php?u= to the WSJ URL and it should work. There's an extension (Bypass Paywalls) for firefox and chrome that does the same thing for a number of sites.
