I suspect this is too little data.
But 10+ deals suggests you have enough inbound interest to test it.
Try AB testing it I would say.
383 karma · joined February 27, 2024
But 10+ deals suggests you have enough inbound interest to test it.
Try AB testing it I would say.
Cries in sadness that my university lab was unable to buy compute from 2020+ when all the interesting research in AI was jumping up and now AI is going into winter finally compute will be cheap again.
I also find economic arguments like you make here generally convincing, arguing against this (that risk increases costs which must be priced in) would be similar to arguing against the Laffer curve. This is essentially close to my position, but different to the person I originally replied to.
I do have an optimistic bent that, through either ignorance, arrogance or risk-seeking, people actually over-risk on experimental therapies, for example, launching Biomed spin offs. Because the personal gain (curing a disease, helping people, or becoming popular) can be so high, it can motivate risk beyond other busiensses. This is probably where I most differ from you: I know many businesses which should not exist, have never turned over profit, and maybe never will. But a rich family, or the savings from a lifetime of earnings sometimes get put into these shots at success. It's not like they get cut once the net profit reaches 0. Same with local shops: A number of the local shops you can't understand running profitably survive because they have paid-off mortgages and only have stocks/staff/licensing to deal with. It makes little financial sense to run them over renting the shop to someone else for more profit, yet people do.
Perhaps this extra reason to stay invested/risk on Biomed is matched by the fraud in that industry, though (Hello Theranos).
Regards the later point, I understood that though my comment could have been phrased better. The meaning was "the government will force [as necessary by asset seizure or bankruptcy] the breach-of-contract party to pay you".
Specifically, once I load something (say a document), if I edit in on another tab, it will re-render on my original tab so long as the document it references (by key) is the same and firebase permissions allow me to read it as that firebase user.
It means all my permissions and schema goes in firebase/firebase rules (e.g. users can only access their own documents of type "Car" but can read all documents of type "Road"). Which is quite handy to do stuff quickly that syncs across users and works as a single source of truth.
It's an add on that works with Svelte (and Firebase). *(sorry, I got the wrong phrase)
It's the fact that options laid out on the table with unknowns can be preferable to people than no options, or fewer options.
The same dichotomy exists for crime vs agreements:
- You'll be compensated if a party to an agreement breaks a clause or costs you money (say they fail to provide $30k of goods to your business)
- But if someone steals your car, the government won't put $30k in your account for a new one.
The same loss occurred. The second case was totally unwarranted, yet the individual outcome is worse.
Why? Because doing it any other way would make things worse. People would lie about stolen goods, conduct insurance fraud, etc.
Unfortunately in some situations, things "which are worse" are a different category and don't have an answer with government intervention.
Say I know about ATP Synthase and how the proteins/molecules involved there interact to make a sort of motor.
How does AlphaFold help us understand that or more complicated systems?
Are proteins quite often dispersed and unique, finding each other to interact with? Or like ATP Synthase are they more of a specific blueprint which tends to arrange in the same way but in different forms?
In other words:
Situation 1) Are there many ATP synthase type situations we find too complex to understand - regular patterns and regular co-occurences of proteins but we don't understand them?
Situation 2) Or is most of the use of Protein situational and one-off? We see proteins only once or twice, very complicated ones, yet they do useful things?
I struggle to situate the problem of Unknown proteins without knowing which of the above two is true (and why?)
For example, early on it says: " collect interesting information, process the information, exfiltrate files, and distribute files, configurations and commands to other systems."
and later on: " they were used, among other things, to collect and process interesting information, to distribute files, configurations, and commands to other systems, and to exfiltrate files."
It also mentions several times that the attack on a South Asian countries embassy was the first time this software was seen.
Repeating info like this was kind of a sign of part-applied AI edits with RAG a while ago, might still be true today.
The try-hard meta seemed to be the opposite of midwit meme e.g.
Newbie: Use the VPN it prevents tracking and password leaks on wifi! Mid level: Don't use the VPN as the VPN owner could be tracking everything! Elite level: Use multiple VPNs for different things, self host Tor and VPNs in other countries
1) Differences from fire-svelte for core functionality?
2) For the beyond core aspects, like Group-WebRTC (great innovation btw), how will things like TURN/PEER servers work to manage the WebRTC connections?
SSDs, Flash, .NET, Dynamic web apps (Web 2.0), HD, Broadband, HTML5(+Websockets), Framework Eco-systems for non-server side code.
When the Will Smith one was released, it was kind of a parody though. Tech had already been able to produce that level of "quality" for about 2 years at the time of it's publishing. The Will Smith one is honestly something you could have created with Disco Diffusion in early 2021, I used to do this back then...
2022 saw: https://makeavideo.studio/ (coherent, but low res - it was possible to upscale at extreme expense) https://sites.research.google/phenaki/ https://lumiere-video.github.io/
It was more like 18-20 months ago sorry so early 2023, but https://runwayml.com/research/gen-1 was getting there as was https://pika.art/home - Sora obviously changed the game, but I would say these two were great.
Must have explained it to 3-4 junior engineers/interns, but could never quite find a link back to it.
Thank you!
2 years ago we had decent video generation for clips
7 months ago we have Sora https://news.ycombinator.com/item?id=39393252 (still silence since then)
With these things, like DALL-E 1 and GPT-3, the original release of the game changer often comes ca. 2 years before people can actually use it. I think that's what we're looking at.
I.e. it's not as fast as you think.
Aider is pretty biased towards Python, for example (it's sample prompts largely use and test on Python)
I semi-liked him at some point, but his video on the T3 stack lost my support. In a few videos he completely ignored that Next.js and other stacks already had shipped the features T3 supposedly uniquely had. T3 is different to work with, in my opinion worse, and he completely ignores the innovation from Next.js that had already happened since the Next.js code he compares it to. It's like he launched a 1995 model car in 2000 and started comparing it to all the 1990s cars.
I know if Next.js hadn't innovated, his framework would have stood a chance at wider immediate adoption. It had some real, and unique pros. Unfortunately, by the time of his release, his ideas and knowledge were either already outdated, or he is hiding the truth from himself. A lot of the supposed benefit of T3 stack is now provided in Next.js by default.
The man should look at usage stats. He's getting more pithy and angry over time. I don't think it's wise.
One thing I wonder, can I ask you? it's quite - prudish - but what do you do about innate motivation and needs?
E.g. when I am not in a relationship, I'm either: - relieving things daily at the same time - sporadically going days without, then multiple times a day etc - withholding from that behavior before dates
These 3 all have pros and cons. I wish I could just withdraw totally, but that's not an option.
I don't understand you can minimize a thing that has an ever growing P over time unless you take some action.
I know eBikes are a massive step up from laptop batteries (seemingly better for the density too somehow?)
There are people in the US who increase their wealth by $1 million each year and pay no income tax because their increase in wealth was tied to assets, not increased income that year. By loss offsetting, or other evasion/avoidance measures, they can even end up paying little or no capital gains/asset taxes (which are already significantly lower than income taxes). And they aren't included in your figure.
These are the same people mentioned in the other comment reply to you, quote "A worker's salary is taxed at a high rate like up to 40%, whereas a wealthy person's gains from assets are taxed at a much lower rate like 15%. And that is if they realize the gains at all, because their unrealized gains don't get taxed at all."
If assets and capital gains were taxed in the same way (no intermingling, no loss offsetting) as income taxes, then tax would be far more equally distributed - which according to you, would be better right?
You think it's unfair to ask "laborers" pay more tax, since the top 1% already pay 45% of income tax
Let's adjust the law then and have the top 5% by wealth pay 45% of all wealth+income increase via tax, this must be fairer to you?
Basically I would agree with your argument if the practical effects were that the people in practice with the most money and secure income(from any form, be it employment or wealth/asset gain) were paying 45% of all the taxes, however that's not the case. And you are ignoring massively the unrealized gains and ability to offset losses which is available only to those with wealth, but not income. (e.g. I can't set aside my gross salary as a net loss and then use an instrument to gift it to my children a few years later with no effective tax)
The current system is designed to A) Easily extract taxes (do it before the commoners get the net money B) Keep assets with those who already own them C) Shift the tax burden on to the top 20-30% of earners by income, but not wealth.
The effective tax rates you quote for the US are both influenced by this system in a way we can't remove from the data, and show anyway that your belief is weaker than the facts you originally presented (ca. the 45% figure which ignored wealth/assets).
However review my last month or so of post history for my views on earned/employment income, unearned income and asset income.
In essence I differentiate between revenue within income, but not income/investment. I don't think taxes should either. Specifically, I think capital gains/intermingling of asset taxes while requiring up to date taxed-upfront employment taxes is one of the worst decisions society ever made.
Part of the reason why we misunderstand other processes in the brain and have since the Lobotomy times enshrined that approach.
We all seek different kinds of quality; I don't find Peruns videos to have any quality except volume. He reads bullet points he has prepared, and makes predictable dad jokes in monotone, re-uses and reruns the same points, icons, slides, etc. Just personally, I find it really samey and some of the reporting has been delayed so much it's entirely detached from the ground by the time he releases. It's a format that allows converting dense information and theory to hour long videos, without examples or intrigue.
Personally, I prefer watching analysis/sitrep updates with the geolocations/clips from the front/strategic analysis which uses more of a presentation (e.g. using icons well and sparingly). Going through several clips from the front and reasoning about offensives, reasons, and locations is seems equally difficult to replicate as Peruns videos, which rely on information density.
I do however love Hardcore history - he adds emotion and intrigue!
I agree with your overall hope for quality and different approaches still remaining stand out from AI generated alternatives.
I just feel intuitively that it's masking the variance by converting it into within-subjects variance arbitrarily.
Here's my layman-ish interpretation:
P-values are easier to obtain when the variance is reduced. But we established P-values and the 0.05 threshold before these techniques. With the new techniques reducing SD, which P-values directly interpret, you need to counteract the reduction in SD of the samples with a harsher P-value in order to obtain the same number of true positive experiments as when P-values were originally proposed. In other words, allowing more experiments to have less variance in group tests and result in more statistical significant if there is an effect size is not necessarily advantageous. Especially if we consider the purpose of statistics and AB testing to be rejecting the null hypothesis, rather than showing significant effect sizes.