Not to mention their customer service is amazing. I travelled several times, been to Shenzen, but mostly travelled to Shanghai for training and procurement of large manufacturing machines. To be honest I am a software person and was never really trained or taught fixing and maintaining moulds/machines.
But they have always been gracious and insanely dedicated to making sure my problems are solved and they show me how to rebuild everything. I have a dedicated WeChat group with their engineers and get back to me 7 days a week.
Compare that with a Candian machine, where I can rarely get a hold of a technican and they lack any deep expertise on anything even just to have a informal conversation on how to improve things at my plant.
Yes, ask any Canadian how they feel about their sovereignty right now. Let me know how the Riviera in Gaza is going to work out and enjoy that vacation in the Gulf of America. Do you remember when Solemeni got assassinated. I also hope your not what he calls "vermin" and end up in Guantanamo Bay. Just because they're so dumb and incompetent and incapable of accomplishing their goals(at this moment) doesn't make a spade a spade. Honestly I am not even surprised by this anymore your post describes exactly why fascism takes root, just endless justification of authoritarianism.
There is no threat towards Canadian sovereignty obviously. The Riviera in Gaza would be a lot better than the current situation in which its biggest industry is exporting rockets into Israel. Also not territorial expansion as Gaza isn't a neighbour. Imperialism? Yes. But look at the world. It functioned much better in most places when imperial powers controlled more of it. That isn't an excuse for them to invade each other, which never ended well. Also he is just using it as a negotiating tactic with Jordan and Egypt obviously.
Changing the name of some water isn't territorial expansion or authoritarianism.
Solemeni was a terrorist. A dead terrorist is a good terrorist. If you attack a sovereign country then you should expect to be targeted. Fuck around? Find out.
Obama closed Gitmo. He promised to, remember, so he must have. And he withdrew from Afghanistan, remember? He said so in between claiming that marriage was only between a man and a woman. What a hero of the left.
Feel free to tell me I am wrong when I actually am instead of just when your imagination runs wild and you see "authoritarianism" in your cereal. Killing terrorists is authoritarianism lol. And drone striking children isn't?
> There is no threat towards Canadian sovereignty obviously. The Riviera in Gaza would be a lot better than the current situation in which its biggest industry is exporting rockets into Israel.
Like there is no threat to Medicaid? Here’s the authoritarianism: Trump is clearly setting us up for a major constitutional crisis by way of his lying bullshit. Go see the way he talked to the governor of maine and tell me you dont hear the threats? You might not be a Nazi, but you’re sure damn close.
It’s not that simple. The brain is a very complex system and once you get other components of your brain to a neurotypical baseline or even better ADHD is a gift.
Those regions being sensory motor cortex and prefrontal cortex. But simply its a system of networks and attention regulation.
Those networks can be trained and I have experienced it first hand by reversing engineering my own brain and training it beyond normal baseline. This training is called neurofeedback and it transformed my ADHD into a super power.
During my evaluation my ability to focus(was measured) was worse than the average ADHD person. Afterwards it was above even normal people.
I need don't even consider myself someone with ADHD and everything just seems to come together naturally. ( Deadlines, managing details, time management and planning- all prefrontal cortex) The networks in my brain have been training to switch normally, and hyper focusing is just focusing but on anything I really want.
It was the cure I’d been searching for my whole life and it actually worked.
Can you divulge more details about your neurofeedback training? What was it like, how long did it take, how did they measure the brain, what stimulus and which "punishment" for bad brain behavior was used?
You have to consider that when you are a new drummer you won't have good control of your power. So you will tend to hit things a lot harder just to keep up. This means you will break sticks and cymbals more often.
I have been playing for 10+ years that was a huge problem for me, but eventually your mind will adapt to dynamic nature of keeping time. I feel most people think keeping time as a singular action, but its mostly managing dynamics and spacing of music. Cymbals are fundamental to accomplish that.
I will say that I personally prefer Sabian since I am a rock/metal drummer. But Zildjian cymbals are thinner and softer they will have less volume but a brighter and more lush sound. Most of their products are designed for softer Jazz and Pop playing. If you want something in between Paiste and Meinl try to fill that gap.
I actually don't hit too hard since I started on a Roland e-drum set and if you hit too hard on those, you risk breaking the pads, so I think I developed better stick control from that. On my acoustic set, I've only ever broken the Zildjian sticks and some cheapo Amazon brand ones. The Vic Firth have held up for over 2 years.
I also use the Sabian AAX line and it works great. I like to play along with rock, hip-hop, and pop. I found the Sabian's to be a good middle ground for now.
Judging by the feedback I'm getting in other comments, I guess the Zildjians are not any less reliable than the Sabians (and I understand the Sabians are derived from Zildjian anyway?). I'll have to give them another go at some point.
Today was the first day I ever used Grafana and what a terrible day its been. Grafana really needs to simplify their interface. There is no point having so many features when you can't even use the most basic functionality without pulling your hair out.
If there is going to be an editable JSON panel at least separate out the presentation and data layers. I don't even know where to start with the query editor, NewRelic allows autocomplete using the text editor so instead of clicking it will use context to figure out what to show. If it weren't were all the available dashboards and pretty looking things I would just use chronograf.
There was a massive sell off(consolidation) happening from Jun 20th until Jul 2 on the S&P. I know I made some profit off a buy position, but noticed that the trend continuation was taking incredibly long time to start, which is rare in a uptrend. This sell off was already in the works, not to mention sell structure that developed at the top. I wouldn't be so quick to say these things are random. Institutions with billions of dollars don't do things randomly. There was no black swan event this time. It was just time for the over bought market not only to correct but to actually reverse.
This is based on institutional theories or smart money of how price plays out in the chart. I have developing my theory for years using those principles(i.e Wycoff Methods[1]
I did not say these things were random at all, my point is that nobody can explain them because of the complexity - when talking about aggregate movements over time. It's like trying to explain the causal chain of events from a buttefly flapping their wings to a hurricane happening across the world. Or that if one of the twins slept closer to the window and saw the outside they will be more creative.
Regarding the technical impact of institutional moves, that is real, if someone does a huge sell block, you can explain the drop by correlating it, but those are isolated events and it's much harder to predict what the market itself will do than to explain, ah, five minutes ago blackrock did their quarterly rebalancing of their ETFs and you can see it in the chart - that's not what these analysts from these articles usually do, they just come up with something that sounds plausible for clicks searching for a reason.
I agree its a concern. You could argue everything on XDA forums and any modified Windows should have the same problems.
Personally I use the community of version of Arch i3, not going to check the source on that, but will trust that community might keep an eye on anything nefarious.
Doesn't seem like Linux is completely immune from that either(i.e. XZ Utils).
But the build is amazingly stable and fast, and I will have to take it good faith that they didn't do anything that would harm me.
I can at least track down the author in various communities and seems to be semi-open about their identity. There is a forum thread any concerns I can bring up with the person responsible.
They are also taking donations. I would hope that exposes them to some liability and financial trail if they are found doing something illegal.
The min/max cost functions make a lot of sense(easiest to visualize). It also can do a lot when you do multi-objective optimization. But I have always wondered how you go about evaluating other cost functions. I mean mathematically the concept is intuitive. i.e. just swap it for a quadratic or exponential, but which cost functions are useful in the real world?
How do you get the output to be formatted correctly or without any branches.
Say for example I want a step-by-step instruction for an action.
But the response will have 1. 2. 3. and sometimes if there are multiple pathways there will long answer with 2.a,b,c,d. This is not ideal I would rather have the most simple case(2.a.) and a short summary for other options. I have described it in the prompt but still cannot get nice clean response without to many variations of the same step.
I have not encountered this problem yet. When I was talking about the format of the answer I meant the following: No matter if you're using Langchain, Llamaindex, something self made, or Instructor (just to get a json back); under the hood there is somewhere the request to the LLM to reply in a structured way, like "answer in the following json format", or "just say 'a', 'b' or 'c'". ChatGPT tends to obey this rather well, most locally running LLMs don't. They answer like:
> Sure my friend, here is your requested json:
> ```
> {
> name: "Daniel",
> age: 47
> }
> ```
Unfortunately, the introductory sentence breaks directly parsing the answer, which means extra coding steps, or tweaking your prompt.
It's pretty easy to force a locally running model to always output valid JSON: when it gives you probabilities for the next tokens, discard all tokens that would result in invalid JSON at that point (basically reverse parsing), and then apply the usual techniques to pick the completion only from the remaining tokens. You can even validate against a JSON schema that way, so long as it is simple enough.
If that's what you need, it would make all sense to redo the instruction fine-tuning of the model, instead of fiddling with prompt or processing to work around the model settings that go counter to what you want.
At the very beginning of my journey I did some fine tuning with Lora on a (I believe) Falcon model, but I haven't looked at it since. My impression was that injecting knowledge via fine tuning doesn't work, but tweaking behavior does. So your answer makes much sense to me. Thanks for bringing that up! I will definitively try that out.
I very much agree with your sentiment. Glad to see someone else recognize the problem with trying to find interesting content nowadays. I have noticed a negative relationship between sizes(after a certain threshold) of communities/platforms and the ability to find interesting content that is also suited to my preferences.
It seems like a ranking problem, you would think having upvotes/likes/points/etc. would be a good metric. But you just get an outsized influence of posts that have huge critical mass, and the content just becomes generic and clickbaity(i.e. Mr. Beast).
I think it is an optimization problem that has to balance competing needs and bring the best solutions under a wide set of criteria. This has been my experience trying to solve this problem, I already have a system that ingests huge amount of reddit posts from preferred and top subreddits. What I want to do is filter out the best content, then apply that generally to all content sources(YouTube, Twitter, TikTok, etc). Moving away from content that gives a short term dopamine hit to content that is an experience that changes structures in your brain. Ideally sustaining pleasure, intrigue, curiosity for longer periods of time.
There are sites with great UX like Scrolller[1] trying to solve the issue, but its only images/gifs.
But they have always been gracious and insanely dedicated to making sure my problems are solved and they show me how to rebuild everything. I have a dedicated WeChat group with their engineers and get back to me 7 days a week.
Compare that with a Candian machine, where I can rarely get a hold of a technican and they lack any deep expertise on anything even just to have a informal conversation on how to improve things at my plant.