_akhe

I saw a RAG demo from a startup that allows you to upload patient's medical docs, then the doctor can ask it questions like:

> what's the patient's bp?

even questions about drugs, histories, interactions, etc. The AI keeps in mind the patient's age and condition in its responses, when recommending things, etc. It reminded me of a time I was at the ER for a rib injury and could see my doctor Wikipedia'ing stuff - couldn't believe they used so much Wikipedia to get their answers. This at least seems like an upgrade from that.

I can imagine the same thing with laws. Preload a city's, county's etc. entire set of laws and for a sentencing, upload a defendant's criminal history report, plea, and other info then the DA/judge/whoever can ask questions to the AI legal advisor just like the doctor does with patient docs.

I mention this because RAG is perfect for these kinds of use cases, where you really can't afford the hallucination - where you need its information to be based on specific cases - specific information.

I used to think AI would replace doctors before nurses, and lawyers before court clerks - now I think it's the other way around. The doctor, the lawyer - like the software engineer - will simply be more powerful than ever and have lower overhead. The lower-down jobs will get eaten, never the knowledge work.

The point is they're backed by the government and military-industrial-complex :) Their annoying and slow "Human Verification" redirect is just an example of how they solve problems - poorly. I would never use that crap and I can't believe anyone on HN does

Who said that? I would ignore them. Live patching just means that your OS updates while it's still running. To answer the question: Raspbian doesn't do that out-of-the-box, you have to restart it after an update (I haven't updated or even restarted the Pi in several months).

Ubuntu and a lot of Linux has the option if you use Ubuntu Pro, but I update my OS so rarely it's a non-issue. I wouldn't even consider that when comparing on-prem vs cloud.

My hunch (and as someone who uses both daily) is that on-prem is 2-10x faster than affordable cloud options within a region, but in some cases like international traffic or high-end multi-core computing etc. cloud should pull away from on-prem - but curious to see what you find :D

React is still the best IMO. It's crazy how much you cover with setState and useEffect alone, I barely use the other hooks. Combined with adjacent tooling and paradigms like JSX, React Native, Next.js, I find the others can't quite keep up with the React ecosystem.

I had to use new Angular on a project recently, it's come a long way and it's actually not too bad especially with GPT-4 explaining the reasoning behind Angular, etc. However, it's no React! It's still just too complex, and I can't be bothered to get into decorators, and other quirks that come along with the Angular world.

If you're liking vanilla JS right now, you might like Node - do some backend stuff. It's just regular JavaScript. I think it's always good to stay brushed up on vanilla JS, I do one-off FE stuff all the time in classic HTML/CSS/JS, but for websites my go-to is definitely still React (Next.js typically). I prefer the declarative style, think it works well for FE state, you end up needing some kind of state management usually.

Sure, curious to see your findings!

An LLM will probably be able to do most of what human minds can do like reason, predict, hypothesize, research, and even get hooked up to other systems to: Visualize, smell, taste, balance, and even direct the movement of limbs, but an LLM can't and won't ever be able to: Feel pain, bliss, anger, sadness, can't feel positive/negative, can't eat/drink, be hungry, feel fatigued, get excited, enjoy things, dislike things, contemplate, meditate, feel warm or cold (though it can detect it), can't feel dizzy (though it can know when it's off balance) - any action where having an experience is a necessary part of what it's doing and the output of it, an LLM is not sufficient to deliver on and never will be.

To compare to a brain, the LLM is like the prefrontal cortex or language and decision network in the outermost layer, but we would still need the amygdala in that metaphor - emotional drives, urges, episodic first-person memories, and experiential components that accompany the language and complete it with personhood.

For raw sensations and tactiles we might need that innermost brain stem - which is probably more chemistry than computation - for the "lights to be on". For example, some jobs will require not just language intelligence, and not just personhood, but for the light behind the images and feelings in the sensations, so that it feels (and would be) alive.

Most if not all of Cloudflare's competitors have been around longer than they have. Their contracts with governments and militaries is why they're valuable, not that nobody knows how to offer a CDN, or other services they sell. They have plenty of competition, and I don't think they lead the market with any of their products.

Amazon started as books, I think it was a reference to how many trees (pages) are in the Amazon.

I like edgelord kinda names: Vercel, Heroku, Mistral, Lightning, Nvidia...

I want my sentences to be like "...so I ejected the vector, ran re-drive, then exported the sequence to Bytestream and now the dyno is back online..."

and less like "...so I pulled the blobs out of Peanut Butter & Jelly and migrated them over to Casserole and you know these days I just serve the dish cold to save on compute"

:D It matters!

This is why I love LLMs. Try using a search engine to find the answer to "how is Cloudflare backed by governments"?

On DuckDuckGo, the entire first page were Cloudflare results! Come on... don't they know what I'm asking?

Let's just get to the answer:

  ollama run llama3

> how is Cloudflare backed by governments?

  Cloudflare has partnerships with various government agencies around the world to help improve internet security, stability, and overall performance. Here are some examples:

  1. *US Department of Defense (DoD): Cloudflare is a certified provider for the DoD's Cybersecurity Maturity Model Certification (CMMC).

  2. *UK Government*: Cloudflare is a partner with the UK Government's Cabinet Office, supporting their efforts to improve internet security and stability in the UK.

  3. *European Union's (EU) Cybersecurity Agency (ENISA)*: Cloudflare collaborates with ENISA, the EU's cybersecurity agency, to help prevent and respond to cyber threats across Europe.

  4. *Australian Signals Directorate (ASD)*: Cloudflare is a partner with ASD, Australia's premier cybersecurity agency, supporting their efforts to improve internet security and stability in Australia.

  5. *Canada's Communications Security Establishment (CSE)*: Cloudflare collaborates with CSE, Canada's national cyber security authority, to help protect Canadian interests online.

  These partnerships often involve:

  1. *Security research*: Cloudflare works with government agencies to identify and mitigate emerging threats, such as DDoS attacks, malware, and other types of cyber attacks.

  2. *Cybersecurity training*: Cloudflare provides training and education to government officials, security personnel, and industry professionals on best practices for cybersecurity and cloud-based services.

  3. *Technology development*: Cloudflare collaborates with governments on developing new technologies and solutions to address specific cybersecurity challenges, such as combating online terrorism or protecting critical infrastructure.

  4. *Incident response*: In the event of a significant cyber incident, Cloudflare works closely with government agencies to respond quickly and effectively, minimizing the impact on affected organizations.

  These partnerships demonstrate the trust and confidence governments have in Cloudflare's ability to help protect their interests online. By  working together, they can create safer, more secure, and more resilient internet ecosystems for everyone.

> how does Cloudflare partner with DoD? Answer in a single paragraph

  Cloudflare partners with the US Department of Defense (DoD) through various initiatives, including Cybersecurity Maturity Model Certification (CMMC), joint cybersecurity initiatives, research and development projects, training and education programs, incident response and recovery efforts, and market research and analysis. Specifically, Cloudflare collaborates with the DoD on threat hunting and cloud security solutions, develops AI-powered tools for detecting and responding to cyber threats, provides secure cloud storage and compute services tailored to the DoD's needs, offers training and education to DoD personnel on best practices for cybersecurity, cloud-based services, and threat hunting techniques, responds quickly and effectively in the event of significant cyber incidents, and analyzes emerging cybersecurity trends and threats to inform the DoD's security strategies and investments.

> how much does dod pay cloudflare

  I couldn't find exact figures on how much the US Department of Defense (DoD) pays Cloudflare, as this information is likely classified or not publicly disclosed. However, I can provide some context:

  1. *Cloudflare's revenue*: In 2020, Cloudflare reported $185 million in revenue, which was a 52% increase from 2019.

  2. *Defense spending*: The DoD has a significant budget for cybersecurity and cloud services. For example, the DoD's 2022 budget request included $15.4 billion for "Cybersecurity" and $3.8 billion for "Cloud Computing Services".

  3. *Cloudflare's defense contracts*: Cloudflare has secured several defense-related contracts with the DoD, including a $40 million contract in 2019 to provide cloud security services to the Defense Information Systems Agency (DISA). Another example is a $20 million contract in 2020 to develop an AI-powered threat detection system for the US Air Force.

I could go on...

What might be scarier is I learned this information from reading regular news articles maybe 5-6 years ago, so I've known for a long time that Cloudflare was vaguely involved with, or at least sold services to the U.S. federal government and military. But in 2024 it's very hard to find anything detailed through Google or DuckDuckGo, or even Wikipedia. You can find it on Cloudflare's website when they want to use it for marketing:

“Cloudflare has long served the most technical, security-aware agencies like the FBI and U.S Department of State,” said Matthew Prince, co-founder & CEO of Cloudflare.

Source: https://cloudflare.net/news/news-details/2022/Cloudflare-Ear...

But I like that the LLM is able to really grab the information in a way search engines are getting worse at (through censorship and other algorithmic pressures).

https://www.exactchange.network is on my Raspberry Pi in my San Francisco apartment.

The front-ends of the sites therein (e.g. https://www.playshadowvane.com/) are all on Vercel, with the backend/APIs running on exactchange on the on-prem Pi.

I can't believe how many people are suggesting Cloudflare who:

1) Is backed by the U.S. Dept of Defense, NIST, the FBI, the Department of Homeland Security, and the EU (Horizon 2020 Program).

2) The main way people know about Cloudflare is that awful Verifying you are human message and 3-second redirect you see on some sites - which is my prompt to immediately leave the site.

Why would anyone want anything to do with slow as hell and government-backed Cloudflare?

How shallow am I if I just don't like the name "Porkbun"?

It's the main reason I never used GoDaddy back in the day, simply didn't want to see or think of that word, just didn't like it lol. Funny that it's the only reason I was never their customer as there was a time their cPanel experience was slightly less bad than the others.

So yeah I'll never use Porkbun and my only shallow reason is their awful brand name.

Sure, like extending i18n to the code itself, I like it. It would be a fun way to practice my Italian.

I've seen code all documented in Japanese with Japanese strings. I thought it looked awesome and wondered why people don't just write code in their own language - why nobody has bothered to create, say, Japanese JavaScript.

There are some non-English based programming languages: https://en.wikipedia.org/wiki/Non-English-based_programming_...

The whole glasses thing might be more like Apple Watch. Feel like it has to be used with the hands to be tool-like enough. Our hands and fingers are so evolved to using things, it's a big part of how we do anything.

Sure there will be smart glasses, but it might be more of a peripheral thing.

That is an incredibly good LLM app idea. Book-to-cinematic. Thousands of years of content ready to go

Literally 1 month ago: Kaiser lays off IT workers for a second time in 4 months

Maybe they kept the wrong ones haha

How many jobs did you get

It's tough in SF. I've been surviving as a contractor for WeWork type solo founders with gaps between contracts. My wife and I started a local business with fairly low overhead last year which has kept me far busier than anything happening in tech, and there's more promise there than in anything I'm doing with software.

Maybe in 2025 with the next president, fresh money, etc.

People do this thing with computers, intelligence, consciousness... where they attribute it to things that clearly don't have those qualities.

If the universe is a computer, I need a new word to describe the tool I'm using - what I normally call "a computer". OTOH, if you call a smartphone a computer, I might have to make some slight modifications to my internal definition, but I can accept that it's technically a computer too - but "the universe"? What are we talking about at this point...

Panpsychism puts forth that "consciousness" is in everything. But that abandons what I mean when I say I'm conscious. Maybe a rock is "conscious", but what I mean by consciousness are things like Donald Hoffman's aroma of coffee, the sound of a drop of water, the feeling of being late to a meeting - those kinds of sensations that wouldn't (couldn't) manifest in a rock.

LLM chatbot companies are calling it "AI" - abandoning the meaning of what we usually mean by intelligence, which is information awareness, not merely its recitation.

Useful conversations nonetheless, but I think different buzzwords could be chosen or even invented to link celestial processes to computational ones - unless there's some agenda, like setting out to prove Simulation Theory or something (which has other problems).

You should have seen how hard Facebook had to fight to control social media, especially on iPhone. They almost lost it during the transition to mobile - it took billions in acquisitions and hiring.

I used to think: "Man why did they turn tech off? There's so much undone, so much opportunity for technology and market disruption!"

But the answer is bubbles. Any time sudden money is made in anything it attracts everyone from everywhere and it immediately becomes corrupted and full of scams and old money institutions. Suddenly people aren't becoming developers to innovate but to become personally financially stable. What started out as mostly uneducated hipsters and hacktivists disrupting and improving is now academia, major corporations, wealthy heirs with their WeWork NFT companies, etc. soaking up what's left of the funds, stagnating the industry, gatekeeping it, and playing a totally different (and highly political) game than we were playing in ~2008-2016.

When the tech world came crashing down in ~2016, at that time there was still a lot to disrupt: Pre-Tik Tok, largely still pre- crypto and AI. SaaS and mobile had reached a peak, and we were ready for something new, but I had no idea what was coming lol - Trump and Hilary and politics, then Covid, and now nobody has jobs like under Bush all over again, it's all politics and it's never been worse for a person's image to identify as a software engineer. This is how it was before it was cool though, nobody wanted to be a developer in ~2003.

But it's a necessary cycle, you can't just keep pumping money endlessly, it gets ridiculous quickly. There has to be periods of on and off and extreme hype cycles to see if something might be, and like a kite or firework some of those take off and impress us, but make no mistake they're all - necessarily - bubbles! Get in while it's hot, get out before it bursts :)

I have to remind you that your original comment was:

  It is bad for a state - at least in relative terms - if it has a good thing and suddenly every state has that good thing.

You were not talking about the FTC, you were talking about the state. Furthermore, the parent comment to which you're replying was:

  It's not bad when people are copying you, it signals that California is ahead of the curve on the issue...

which was a reply to

  I suspect this is bad for California

We were all - and especially you - very clearly talking about whether or not it's bad for California. Only later in the conversation did you start to shift your statements to seem as if you're talking about the FTC.

Moving the goalpost fallacy:

  Moving the goalposts is an informal fallacy in which evidence presented in response to a specific claim is dismissed and some other evidence is demanded.

e.g. "It might be bad for California, but not the FTC", "you can't read", "incorrectly inferring", "assuming" etc. Nope, none of those, just sticking to the topic.

Really cool! I like that they have live demos to prove it out. Thanks for sharing

Ahh got it, so not for a real-time consumer use cases then.

Typically when I think of a user using an AI product they're hitting the LLM a lot and expecting results on their screen right then, but I guess there are a lot of back office and data processing use cases.

Just links to a login/signup form. A "Show HN" IMO should be usable or at least viewable.

Yeah I know of Terraform (for me it was via AWS) but I just wonder how it's that valuable. For personal use, I never drank the koolaid on IaaS to begin with. Always found PaaS to be a nicer experience and I like that it actually simplifies DevOps, doesn't add complexity like an AWS or GCP does. I figure if I want more control over the server I can just use a Linux on-prem (no cost) or virtual server and I can fully control the machine - where IaaS like AWS/GCP just feel like expensive jargon hell with too many products. For a larger org sure, you need regional deployments, IAM, and some other stuff - but mostly stuff that is peripheral to code and its hardware requirements.

My favorite DevOps setup is my Raspberry Pi home server running Raspbian, love this thing - WiFi, touch screen so I can hold it like a mobile device or just set it down somewhere while it's serving several APIs, websites, etc. all the time including a local business in SF. Haven't stopped or restarted it in months.

I look at some of these big, old behemoths, and just don't get it. Take Oracle - when you really get into what they "do" it's like... oh... so, a database? Right now they offer clone services of the other cloud providers too, and some other things, but it's mostly just those huge consulting contracts. I just wonder how they get them (and at those rates) if not for relationships, it doesn't seem like their technology is particularly good.

Personally I run stuff like React sites on Vercel, backends on a mix of my Raspberry Pi and Heroku, and 1 thing still in GCP that I can't wait to port out of there. Still looking for a home for my LLMs. As an individual developer, I will probably embrace PaaS and convenience more and more with regards to DevOps, but yeah interesting to see where open-source Terraform goes - would be cool to see companies doing more customized infra internally instead of everyone using AWS.

Question: Is the tldr of companies like these that they sell enterprise server software? And often own the hardware too (data centers)? And then sell a bunch of consulting services on top of that to Fortune 500s and governments? It's tempting to think "How are these guys even relevant anymore?" but IBM's making $60B+ a year with over $10B cash on hand, apparently from mostly "consulting services".

For a lot of developers including me, I never think about IBM or HashiCorp (or Oracle, SAP, etc.) and it's hard to imagine why someone would want to use their software compared to something newer, friendlier, cheaper, and probably faster. Is it just relationships?

Just curious how customers are actually getting value from an IBM or a HashiCorp or an Oracle.

Most hilarious attempt at moving goalposts I've seen all week - this conversation has been about whether or not its bad for California the whole time, now you're switching to say whether or not its good for the FTC! Just funny

> This is why Google has the technical ability to scan your email content, and Proton doesn't

I don't buy this theatre for 1 second. Proton obviously has the ability to run a data business off my data if they wanted to - they own it all and it runs on their server.

Not saying they do it, but the idea that it's impossible because of "encryption" is a joke - you assign the keys and you could easily use them on the backend out of my view, the user has no visibility into how that works or what you guys do behind the scenes - it's not like decentralized technology (even the best decentralized tech can't deliver on this promise yet).

So it seems like it's slow because of client-side encryption which Google also does. Yes I know they have a data business, but the data is still encrypted.