I noticed this interesting fact during my secondary school education. Additionally you can use this to prove that e is the point where a change in the exponent causes a larger result than a change in the base. E.g. (x+d)^y > x^(y+d) when y >= e for any real value d.
We (and many others) have a team building fun things into our data analysis tool here.
For what will soon become a 10% time prompt engineering role for a much easier kind of security investigations experience, we are hiring cleared security folks in Australia (SIEM / python SE) and a cleared cybersecurity data scientist in the US. See Google docs @ graphistry.com/careers
Likewise, if you use a SIEM/Splunk/Neo4j/SQL today and want a better experience for it, feel free to ping for the early access program. You can see our Nvidia GTC talk on the GPU SOC for types of experiences we are building in general. GPT 3 already enabled way easier experiences here, and then GPT 4's quality jump shifted it from feeling working with a weirdly well-read 10yr old to working more with a serious colleague.
Serious question. How can you reconcile needing CLEARED individuals to perform the work but give the data to a non-cleared entity who seems to have issues with security?
Perhaps there’s now a self hosted or enterprise version where they promise not to leak it?
We work with everyone from individual university researchers trying to understand cancer genomes or European economic plans in their graph DBs, to big corporations struggling with supply chains in Databricks, to government cyber & fraud teams using Splunk. For many, an OpenAI/Azure LLM is fine, or with specific guard rails they've been having us put in.
But yes, when talking with banking & government teams, the conversation is generally more around self-hosted models. Privacy + cost both important there -- there is a LOT of data folks want to push through LLM embeddings, graph neural nets, etc. We generally prefer bigger contracts in the air-gapped-everything world, especially for truly massive data, though thankfully, costs are plummeting for LLMs. Alpaca/Dolly are great examples here. Some folks will buy 8-100 GPUs at a time, so this is no different for those. My $ is on continuing to shrink LLMs down to regular single-GPU being fine for many scenarios. The quality jump of GPT4 has been amazing, so it's use case dependent: data cleaning seems fine on smaller models, while we love GPT4 for deeper analyst enablement. Wait 6mo and it's clear there'll be ~OSS GPT4, and for now, even GPT3.5 equivs via Alpaca-style techniques are interesting, a lot of $ has begun moving around.
LLM side is new from a use case perspective but not as much from an AI sw/hw pipeline view. Just "a bigger bert model". A lot of discussions with folks has been extrapolating with them based on what they're already doing with GPUs, where it's just another big GPU model use case. Internally to us, as product team doing a lot of data analyst UX & always-on GPU AI pipeline work... a very different story, its made what was already a crazy quarter even that much more nuts.
QOI caused a huge burst in interest for image formats. Audio is a little harder to work with and understanding so I can imagine less impact for QOA (still good though). What I'm really looking forward to is QOV (video).
The physical layout, sure, but would it conceptually involve more than just defining twice as large chunk types for 16 bit channels? Unless I'm missing something, it should be fairly uncomplicated code-wise?
The problem with a QOI that would have 16-bit is that lossless becomes more expensive, exact match in the color table is more rare too and not worth it anymore.
You will start to need more prediction modes + offsets.
Not really. QOI's opcodes are very specifically designed based on the characteristics of 8 bit images. For example, there is an 8 bit opcode for a run of up to 64 identical pixels. This makes a lot of sense for an 8 bit image, but is a lot less likely to appear in a 16 bit picture. For a 16 bit photo, the minimal QOI modification would almost certainly be worse than uncompressed since noise in the sensor will prevent any of the 8 bit opcodes from working.
RLE doesn't really work for photos even at 8 bit depth, the noise is noticeable even there. It's mainly useful for drawn/rendered content, where solid colored areas are far more common. That won't be much different at 16 bit depth.
A possibly better example for photos would be the 8 bit tag that says a pixel is within -2 to +1 in each of RGB. In 8 bit, there are a lot of cases where this is a big enough range to capture the noise of a static color (and or a gradient). In a 16 bit image, a lot of the time there might be 3-4 bits of noise which will mean that the tag is a lot less useful.
I render accumulation fractals like the Buddhabrot at 32 bits per channel. So far I have only used about 27 bits but it's nice to have a bit of headroom.
Maybe by asking the AI to be concise they're reducing the server load? It has been running awfully slow recently, probably due to a surge in mainstream awareness.
I'm not surprised, back when I was using them at uni they turned off my mobile data without warning.
Called them multiple times, no-one would tell me why. Went to a Three shop and they told me I'd been disconncted for "security" because I'd been away from the address on my account too long!
Plus their broken app and abysmal coverage in London...
In their examples section they mention how a double agent was charged with tax evasion based on $2m of undeclared income from bribes.
But if the agent WERE to mention the income on their tax documents, even as "Unspecified due to 5th amendment protection" that surely would be incriminating enough to cause further investigation from the state. (I assume the US government keeps a close eye on their agents papers).
I'm just not sure how you're supposed to declare ill-gotten gains without 5th amendment worthy self-incrimination.
That’s why they amended the constitution to collect income taxes. The 16th amendment was necessary for a variety of reasons.
The Supreme Court - the constitutional court - kept squashing attempts to collect one.
It might not be important that it undermines your 5th amendment right when the 16th amendment allows any means necessary - or doesnt limit collection methods.
If the goal of the IRS is to collect as much tax money as possible, they would, in such a case, have all the incentives to protect your illegal business from being busted.
The IRS couldn't manage to keep the detailed tax records of every american out of propublica's hands... Does anyone think they can keep them out of other agencies hands?
I would imagine that in cases where you're earning a significant enough amount of money, the issue isn't whether the government knows you're up to no good, the issue is whether they have admissable evidence that you're up to no good.
My contention is that the mere knowledge that you are reporting income is actionable information to the government and the IRS has no way to guarantee it does not get out.
We should work under general provision: if data is gathered, it will be used. Even if it is not legal today for IRS to do something, it does not mean it will not be tomorrow. And law unfortunately works backwards too.
They don't not share them out of the goodness of their hearts. They're so intentionally understaffed and overworked that they probably don't have time to refer all but the most egregious cases.
"The United States Supreme Court has ruled in Haynes v. United States that the Fifth Amendment to the United States Constitution exempts felons—and, by extrapolation, all other prohibited possessors—from the registration requirements of the Act. "
Would someone pleading the 5th count as probable cause? I would hope not. That would go against the idea that you can't be punished for exercising a right.
Go on. How is forcing someone to testify against themselves with the threat of criminal charges (tax fraud!) and then using that information in a criminal investigation not a 5th amendment violation.
I’m not a lawyer, I don’t know anything. I’m just curious.
I believe that the current legal situation goes further than just keeping people from being compelled to testify. Pleading the 5th also can't be used as evidence against someone (because that would be a form of compulsion), or even as probable cause.
I asserted that 5th amendment should prevent the government from using this coerced testimony.
Would this not taint any investigation started on the basis of coerced testimony? Why would the exclusionary rule not kick in here?
As far as I understand in a criminal context no adverse inference may be drawn from ones refusal to self-incriminate. How would investigating someone for refusing to self-incriminate to the IRS not be exactly that?
>I asserted that 5th amendment should prevent the government from using this coerced testimony.
Again, you did no such thing:
> Presumably investigating someone over that would violate their 5th amendment rights, no?
An investigation can be initiated, and can proceed, without any testimony whatsoever from the accused.
Which coerced testimony are you attempting to refer to? The cops show up and ask you about your tax return, and you say "". Either they continue to investigate you or they don't, but you have not been coerced into saying anything at all.
> Presumably investigating someone over that would violate their 5th amendment rights, no?
>I asserted that 5th amendment should prevent the government from using this coerced testimony.
This kind of revisionist interpretation, that a poster will engage in to make an altogether different assertion, is not uncommon on HN. Pointing it out gets downvoted, overall making the conversations worse because it's viewed as nitpicking, when it's really just trying to avoid bad faith...wasn't there an article posted about this recently?
If the government asks if you, Bill the graffiti artist, painted any graffiti on the wall of a giant chicken, and you said no. They could still apply a penalty for lying to an LEO if that exists as a law if they can catch you later.
What if everyone included "I plead the fifth" in their tax filings? Perhaps as a form of protest, or because CFAs and distributors of tax-filing software started adding it to their boilerplate?
The performance hit is quite unacceptable in games such as CSGO. I cannot really think of any competitive game where the performance hit is justifiable.
It does seem possible, e.g. rather than sending an uppercase alphabetic message as 8-bit ASCII bytes, you can send their position in the alphabet instead, so 01000001 becomes 00001.
Of course this only works when you're limiting yourself to a subset of all possible data (in this case, the uppercase alphabet). A general method where all possible binary combinations are indexed (and the indexes are sent rather than the data) would not save any data as the binary data would always be its own index.
