spacemanspiff01

Yes, but the fact that it is even a question is absurd.

I guess my thought is that you have a boss that told you to touch a computer system, normally only a few people touch it, because of you make a mistake, your company looses 100 billion dollars.

You make pretty sure that whoever is dealing with that system are experts, you make sure that they do not modify it without following procedures, you double/triple check that their change won't have side affects, you then run the changes on a test system just to make sure.

And yeah, you are only adding a print statement, it should be fine...

But what if it's not.

Move fast and break things is fine if worst case scenario your social media site goes down.

Move fast and break things is less fine when you could fuck up a key piece of financial infrastructure.

I don't have evidence of anything, no one does. We don't have evidence that there won't be any mistakes either.

In the past we could expect reliability, because the system was always reliable, within the processes and systems that people implemented around it.

It seems to me that a lot of those processes have been skipped.

How about money in the financial system. What would happen if someone makes a mistake, and something causes bond payments to fail?

It is an outside risk, but based on the reporting it seems like they are digging into code of large complex financial systems...

How much do you trust 1 guy to not make _any_ mistakes.

Maybe this is wrong and they are only looking at read only access of the payment system, but it seems sketchy that this is even a risk.

Or a proposal/feedback process. Ala you are hired by non technical person to build something, you generate requirements and a proposed solution. You then propose that solution, they give feedback.

Having a feedback loop is the only way viable for this. Sure, the client could give you a book on what they want, but often people do not know their edge cases, what issues may arise/etc.

My impression was that most opposition to the jones act was that it has not really worked, and as a side effect drastically reduced the intra-us shipping industry.

Without the jones act, we would not have a ship building boom, but we probably would have more intra-US shipping via ship/rivers/etc.

The counter to this is that it might make more sense for China to _not_ attack the United States with the anticipation that they sit it out. China attacking US forces/ naval bases makes it much harder for a president to sit back and say not our problem/ focus on economic sanctions.

Imagine if Russia started the invasion of Ukraine by bombing polish railways, so that the Ukrainians would not be able to get supplies/resources from the EU. I would think that the EU/Nato response to that would be much more severe than what happened in reality.

While Guam might be considered different, as most Americans cannot place it on a map and it is on the other side of the world, seeing caskets of all the US troops dying makes it pretty hard to politically shrug off as not our problem.

Any idea how many channels this is? Just says multi-channel...

From the laws text:

For the purposes of this Regulation, the following definitions apply:

(1) ‘AI system’ means a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments; Related: Recital 12

https://artificialintelligenceact.eu/article/3/ https://artificialintelligenceact.eu/recital/12/ So, it seems like yes, software, if it is non-deterministic enough would qualify. My impression is that software that simply takes "if your income is below this threshold, we deny you a credit card." Would be fine, but somewhere along the line when your decision tree grows large enough, it probably changes.

...

What is a cross platform sandboxing technology for native apps?

Oh, they will deploy the racks at customers sites?

I thought cerebras had moved to a cloud model so that they could more easily manage/patch their systems?

Can't comment on the bad teeth, but I don't think iodine is causing your c++ issues...

What companies allow you to do work on a personal computer? That seems like a recipe for disaster...

Where I find coding assistants the most useful _is_ in writing code that I already want to write.

Ala - I need to write this unit test, it has these checks, it validates these methods.

Or write a log message for me about what error got encountered here. Those are annoying to write out, but often the llm has enough context that I just start to write and it completes it appropriately.

All of these are things I can easily do myself, are easy to validate correctness, but if I were to write them would consume my limited mental energy for the day.

The other side of this is that if this is over investment (likely)

Then in 5 years time resources will be much cheaper and spur alot of exploration developments. There are many people with many ideas, and a lot of them are just lacking compute to attempt them.

My back of mind thought is that worst case it will be like how the US overbuilt fiber in the 90s, which led the way for cloud, network and such in 2000s.

I guess more generally money laundering.

With regards to bank demobilization here is an interesting article.

https://www.bitsaboutmoney.com/archive/debanking-and-debunki...

But the only way to do this is if it is trained on how to map the word token to character tokens ie

Hello -> h e l l o 66547 -> 12 66 88 88 3

Or, maybe it memorized that hello has a single e.

Either way, This seems to be a edge case that may or may not exist in the training data, but seems orthogonal to 'reasoning'

A better test case would be how it performs if you give the spelling mappings for each word the context?

That's a tokenizer issue though?

How does 174 work with a outside contractor company?

Why would you be unemployable?

what do you think about these guys: https://exa.ai/

Yes, especially for something creative and collaborative like engineering. Creating a culture where remote work works as well as in person is hard, and takes explicit effort and buying from management.

Maybe I am cheap, but I have been using bing because of their rewards points stuff, at least then I get paid for my data.

Same, I can't believe I spent that much time, the guided meditation killed me.

Northern Europe seems to do recycling well, there is a 25 cent bounty on each bottle, and grocery stores have automatic sorter/collectors.

Would there be a way to do the stats gathering on device, then once every few months send a popup with statistics?

Not sure what bias it adds

Like

"hey, we make this app, and we care about privacy, here is the information we have gathered over your usage for the past month, can we send this to ourselves, so that we can use it to improve the app?"

And then show human readable form of what data was collected.

I wonder if these are tokenization issues? I really am curious about metas byte tokenization scheme...

Except it is explicitly not private?

So what I got (but I don't really know what I am talking about)

When thinking about a story (or a sequence of events) we can parse it in different ways.

We may think about the changes in time/location:

-------

1 he went to the store

2. then went to gas station

3. then went home

4. the next day he woke up.

---------

Or you could think of the emotional changes:

---------

1. he was concerned about being unemployed.

2. then at the gas station he got a call saying he got the job and he starts tomorrow.

3. He is nervous about his first day of work.

------

With fmri we can detect these context shifts.

The thing the paper adds is that by prompting the user to pay attention to 'emotions' or 'location' it affects where these segmentation changes occur in the fmri results.