I think it's ability to consume information is one of the scarier aspects of AI. NSA, other government, and multi-national corporations have years of our individual browsing and consumption patterns. What happens when AI is analyzing all of that information exponentially faster than any human code and communicating with relevant parties for their own benefit, to predict or manipulate behavior, build psychological profiles, identify vulnerabilities, etc.<p>It's incredibly amusing to me reading some people's comments here critical of AI, that if you didn't know any better, might make you think that AI is a worthless technology.
All hype and thought experiments about superintelligence and open questions about creativity and learning and IP aside, this is the area that gives me the biggest pause.<p>We've effectively created a panopticon in recent years—there are cameras absolutely everywhere. Despite that, though, the effort to actually <i>do</i> something with all of those feeds has provided a sort of natural barrier to overreach: it'd be effectively impossible to have people constantly watching all of the millions of camera feeds available in a modern city and flagging things, but AI certainly could.<p>Right now the compute for that is a barrier, but it would surprise me if we <i>don't</i> see cameras (which currently offer a variety of fairly basic computer vision "AI" alerting features for motion and object detection) coming with free-text prompts to trigger alerts. "Alert me if you see a red Nissan drive past the house.", "Alert me if you see a neighbor letting his dog poop in my yard.", "Alert the police if you see crime taking place [default on, opt out required]."
> What happens when AI is analyzing all of that information...<p>They run simulations against N million personality models, accurately predicting the outcome of any news story/event/stimulus. They use this power to shape national and global events to their own ends. This is what privacy and digital sovereignty advocates have been warning the public about for over a decade, to no avail.
Models are mediocre solo consumers: they skim, paraphrase and confidently miss the one subtle thing that actually matters. Humans are still better at deciding which three paragraphs in a 40‑page spec are load‑bearing. But as soon as you treat the model as a stochastic code monkey with a compiler, test suite, linter and some static tooling strapped to its back, it suddenly looks a lot more like “creation with a very fast feedback loop” than “consumption at scale”.<p>The interesting leverage isn’t that AI can read more stuff than you; it’s that you can cheaply instrument your system (tests, properties, contracts, little spec fragments) and then let the model grind through iterations until something passes all of that. That just shifts the hard work back where it’s always been: choosing what to assert about the world. The tokens and the code are the easy part now.<p>This might make it into this week's <a href="https://hackernewsai.com/" rel="nofollow">https://hackernewsai.com/</a> newsletter.
Don't forget guardrails and other tweaks you don't know are being applied. I was exploring energy usage and when I reached solar energy somehow the AI decided it was political and switched to useless mode until I explicitly told it to look back at the conversation and it's context and that I wasn't trying to get it to say solar was good and it was OK if numbers made solar look good. I was really weird.
.. except you are completely wrong in a certain way.. AFAIK Google has been reading and indexing patent applications and maybe SEC filings since before word2vec. Certain niches absolutely are reading documents faster than your attorneys can...
I often see things like this and get a little bit of FOMO because I'd love to see what I can get out of this but I'm just not willing to upload all these private documents of mine to other people's computers where they're likely to be stored for training or advertising purposes.<p>How are you guys dealing with this risk? I'm sure on this site nobody is naive to the potential harms of tech, but if you're able to articulate how you've figured out that the risk is worth the benefits to you I'd love to hear it. I don't think I'm being to cynical to wait for either local LLMs to get good or for me to be able to afford expensive GPUs for current local LLMs, but maybe I should be time-discounting a bit harder?<p>I'm happy to elaborate on why I find it dangerous, too, if this is too vague. Just really would like to have a more nuanced opinion here.
I've been analyzing my Obsidian vault using local LLMs that I run via Apple's mlx_lm. I'm on an M4 MacBook Pro with 48GB RAM.<p>The results are ... okay. The biggest problem is that I can't run some of the largest models on my hardware. The ones I'm running (mostly Qwen 3 at different numbers of parameters and quantization levels) often produce hallucinations. Overall, I can't say this is a practical or useful setup, but I'm just playing around so I don't mind.<p>That said, I doubt SOTA models would be that much better at this task. IMO LLM generated summaries and insights are never very good or useful. They're fine for assessing whether a particular text is worth reading, but they often extract the wrong information, or miss some critical information, or over-focus on one specific part of the text.
> I'm just not willing to upload all these private documents of mine to other people's computers where they're likely to be stored for training or advertising purposes.<p>And rightfully so. I've been looking at local LLMs because of that and they are slowly getting there. They will not be as "smart" as the big models, but even a 30B model (which you can easily run on a modern Macbook!) can do some summarization.<p>I just hope software for this will start getting better, because at the moment there is a plethora of apps, none of which are easy to use or even work with a larger number of documents.
The docs I upload are ones I'd be OK getting leaked. That also includes code. Even more broadly, it also includes whatever pics I put onto social media, including chat groups like Telegram.<p>This does mean that, useful as e.g. Claude Code is, for any business with NDA-type obligations, I don't think I could recommend it over a locally hosted model, even though the machine needed to run a decent local model might cost €10k (with current price increases due to demand exceeding supply), that the machine is still slower than what hosts the hosted models, that the rapid rate of improvement means a 3-month delay between SOTA in open-weights and private-weights is enough to matter*.<p>But until then? If I'm vibe coding a video game I'd give away for free anyway, or copy-editing a blog post that's public anyway, or using it to help with some short stories that I'd never be able to charge money for, or uploading pictures of the plants in my garden right by the public road… that's fine.<p>* When the music (money for training) stops, it could be just about any provider whose model is best, whatever that is is likely to still get distilled down fairly cheaply and/or some 3-month-old open-weights model is likely to get fine-tuned for each task fairly cheaply; independently of this, without the hyper-scalers the supply chains may shift back from DCs to PCs and make local models much more affordable.
It really depends on how deep you want to go. And this will likely not be useful in any way, other than a new hobby. Me and my friends who do this kind of thing, we do it for fun.<p>If it was not fun for me, I would not have bought 3 GPUs just to run better local LLMs. Actual time, effort and money spent on my local setup compared to the value I get does not justify it at all. For 99% of the things I do I could have just used an API and paid like $17 in total. Though it would not have been as fun. For the other 1% I could have just rented some machine in cloud and ran LLMs there.<p>If you don't have your private crypto keys in your notes worth millions, but still worry about your privacy, I'd recommend just renting a machine/GPU in a smaller cloud provider (not the big 3 or 5) and do these kind of things there.
I don't really buy this post. LLMs are still pretty weak at long contexts and asking them to find some patterns in data usually leads to very superficial results.
No one said you cannot run LLMs with the same task more than once. For my local tooling, I usually use the process of "Do X with previously accumulated results, add new results if they come up, otherwise reply with just Y" and then you put that into a loop until LLM signals it's done. Software-wise, you could add so it continues beyond that too, for extra assurance.<p>In general for chat platforms you're right though, uploading/copy-pasting long documents and asking the LLM to find not one, but multiple needles in a haystack tend to give you really poor results. You need a workflow/process for getting accuracy for those sort of tasks.
OP's post is trying to con you into using a LLMs for a task that they do not perform well at.<p>This is specific, but if you start replying to LLM summaries of emails, instead of reading and responding to the content of the email itself, you are quickly going to become a burden socially.<p>The people you are responding to __will__ be able to tell, and will dislike you for your lack of consideration.
> No human could read all of this in a lifetime. AI consumes it in seconds.<p>And therefore it's impossible to test the accuracy if it's consuming your own data. AI can hallucinate on any data you feed it, and it's been proven that it doesn't summarize, but rather abridges and abbreviates data.<p>In the authors example<p>> "What patterns emerge from my last 50 one-on-ones?" AI found that performance issues always preceded tool complaints by 2-3 weeks. I'd never connected those dots.<p>Maybe that's a pattern from 50 one-on-ones. Or maybe it's only in the first two and the last one.<p>I'd be wary of using AI to summarize like this and expecting accurate insights
> it's been proven that it doesn't summarize, but rather abridges and abbreviates data<p>Do you have more resources on that? I'd love to read about the methodology.<p>> And therefore it's impossible to test the accuracy if it's consuming your own data.<p>Isn't it only if it's hard to verify the result? If it's a result that's hard to produce but easy to verify, a class which many problems fall into, you'd just need to look at the synthetized results.<p>If you ask it "given these arbitrary metrics, what is the best business plan for my company?" It'd be really hard to verify the result. I'd be hard to verify the result from anyone for that matter, even specialists.<p>So I think it's less about expecting the LLM to do autonomous work and more about using LLMs to more efficiently help you search the latent space for interesting correlations, so that you and not the LLM come up with the insights.
Similar to P/NP, verification can often be faster than solving. For example, you can then ask the AI to give you the list of tool complaints and the performance issues. Then a text search can easily validate the claim.
I think as long as you keep a skeptical loop and force the model to cite or surface raw notes, it can still be useful without being blindly trusted
> “I'd be wary of using AI to summarize like this and expecting accurate insights.”<p>Sure, but when do you have accurate results when using an iterative process? It can happen at the beginning or at the end when you’re bored, or have exhausted your powers of interrogation. Nevertheless, your reasoning will tell you if the AI result is good, great, acceptable, or trash.<p>For example, you can ask Chat—Summarize all 50 with names, dates and 2-3 sentence summaries and 2-3 pull quotes. Which can be sufficient to jog your memory, and therefore validate or invalidate the Chat conclusion.<p>That’s the tool, and its accuracy is still TBD. I for one am not ready to blindly trust our AI overlords, but darn if a talking dog isn’t worth my time if it can make an argument with me.
> ...and it's been proven that it doesn't summarize, but rather abridges and abbreviates data.<p>I don't really know what this means, or if the distinction is meaningful for the majority of cases.
Your colleagues using the tech will be far ahead of you soon, if they aren’t already.
... I mean, what tools one is supposed to be using, according to the advocates, seems to completely change every six months (in particular, the goto excuse when it doesn't work well is "oh, you used foo You should have used bar which came out three weeks ago!", so I'm not sure that _experience_ is particularly valuable if these things ever turn out to be particularly useful.
Far ahead in producing bugs, far ahead in losing their skills, far ahead in becoming irrelevant, far ahead in being unable to think critically, that's absolutely right.
The new tools have sets of problems they are very good at, sets they are very bad at and they are generally mediocre at everything else. Learning those lessons isn’t easy, takes time, and will produce bugs. If you aren’t making those mistakes now with everyone else, you’ll be doing them later when you do decide to start catching up and it will be more noticeable then.
Disagree. For the tools to become really useful (and fulfill the expectations of the people funding them) they will need to produce good results without demanding years of experience understanding their foibles and shortcomings.
And all of those things (good at, bad at, the lessons learned on current models current implementation) can change arbitrarily with model changes, nudges, guardrails, etc. Not sure that outsourcing your skillset on the current foundation of sand is long term smart, even if it's great for a couple of months.<p>It may be those un-learning the previous iteration interactions once something stable arrives that are at a disadvantage?
"The market can stay irrational longer than you can stay solvent" feels relevant here.
We know from the era of data the power of JOIN. Bring in two different data sources about a thing and you <i>could</i> produce an insight neither of them could have provided alone.<p>LLMs can be thought as one big stochastic JOIN. The new insight capabilities - thanks to their massive recall - is there. The problem is the stochasticity. They can retrieve stuff from the depths and slap them together but in these use cases we have no clue how <i>relevant</i> their inner ranking results or intermediary representations were. Even with the best read of user intent they can only <i>simulate</i> relevance, not really compute it in a grounded and groundable way.<p>So I take such automatic insight generation tasks with a massive grain of salt. Their simulation is amusing and <i>feels</i> relevant but so does a fortune teller doing a mostly cold read with some facts sprinkled in.<p>> → I solve problems faster by finding similar past situations → I make better decisions by accessing forgotten context → I see patterns that were invisible when scattered across time<p>All of which makes me skeptical of this claim. I have no doubt they feel productive but it might just as well be a part of that simulation, with all the biases, blind spots etc originating from the machine. Which could be worse than not having used the tool. Not having augmented recall is OK, forgetting things are OK - because memory is not a passive reservoir of data but an active reranker of relevance.<p>LLMs can’t be the final source of insight and wisdom, they are at best sophists, or as Terrence Tao put it more kindly, a mere source of cleverness. In this, they can just as well augment our self-deception capacity, maybe even more than counterbalancing them.<p>Exercise: whatever amusing insight a machine produces for you, ask for a very strong counter to it. You might be equally amused.
The article is more about offloading your thinking to the machine than a real usage of what notes is. You may as well make every decision rely on a coin toss.<p>I take notes for remembrance and relevance (what is interesting for me). But linking concepts is all my thinking. Doing whatever rhe article is prescribing is like sending someone on a tourist trip to take pictures and then bragging that you visited the country. While knowing that some pictures are photoshopped.
I disagree, what ai brings to thr table is instant and total recall of our thoughts/notes/experiences. Deep analysis of that vast data storage is only possible via ai, which should trigger the aha!!! Moment, or the "you're crazy ai" moment. Either way, it's very useful. And we haven't even talked about the knowl she we have collecting digital dust in emails, and notes and reports of past employees.
There's a lot more dimension to the notes we take than what is actually written down. You can share your notes to other people and their interpretation would be very different to what you intended. It even happens with full books and articles. Even lots of metadata don't really help.<p>Text is a very linear medium. It's just the spark while our wealth of experiences is the fuel. No amount of wrangling the word "pain" will compare to actually experiencing it.<p>You'll better be served by just having a space repetition system for the notes you've taken. In this way, you'll be reminded of the whole experience when you took the note instead of reading words that were never written by someone who have lived.
AI at its best can be improvement on human recall, but even at its best it is not<p>> instant and total recall of our thoughts/notes/experiences<p>Closest is with vector searches & RAG etc., but even that isn't total recall because it will misclassify stuff with current SOTA.<p>Throwing everything in a pile and hoping an LLM will sort it all out for you, is at present even more limited.<p>They're good, sure, but you're overstating them.
Hrm. Pretty much every LLM summary I've seen of a document I've read, or a meeting I've attended, at absolute best misses important details and overemphasises trivia, and often just flat-out makes stuff up. I'm not sure I want my own _thoughts_ filtered through that.
How is that different from using selecting a random portion of the archive? If all you want is a sudden remembrance that you maybe have forgotten, you don't need an AI to do that. It's just another complete irrelevant use of AI.
[dead]
At least half of AI's "superpower" in OP's case is the fact that he has everything in Obsidian already. With all of that background context, any tool becomes super valuable in evaluating & guiding future actions.<p>Still, all credit to him for creating that asset in the first place.
Agree with OP that LLMs are a great tool for this use case. It's made possible because OP diligently created useful input data. Unfortunately OP's conclusion goes against the AI hype machine. If "consuming" is the "superpower" of AI, then the current level of investment/attention would not be justified.
Wrt temperature/randomization is it not possible for it to create something genuine. Even life there seems to always be some inspiration for things being made. How did Tesla go from Brushed to Brushless AC motors. There was some foundational knowledge of electricity similar to airplanes, the Wright Brothers their airplane seems backwards (like a canard) but still a plane/needs wings. Not something radical like ion engines for lift (much harder).
I was in a research math lecture the other day, and the speaker used some obscure technical terminology I didn't know. So I dug out my phone and googled it.<p>The AI summary at the top was surprisingly good! Of course, the AI isn't doing anything original; instead, it created a summary of whatever written material is already out there. Which is exactly what I wanted.
My counterpoint to this is, if someone cannot verify the validity of the summary then is it truly a summary? And what would the end result be if the vast majority of people opted to adopt or deny a position based on the summary written by a third party?<p>This isn't strictly a case against AI, just a case that we have a contradiction on the definition of "well informed". We value over-consumption, to the point where we see learning 3 things in 5 minutes as better than learning 1 thing in 5 minutes, even if that means being fully unable to defend or counterpoint what we just read.<p>I'm speficially referring to what you said: "the speaker used some obscure technical terminology I didn't know" this is due to lack of assumed background knowledge, which makes it hard to verify a summary on your own.
At least with pre-AI search, the info is provided with a source. So there is a small level of reputation that can be considered. With AI, it's a black box that someone decides what to train it on, and as someone said elsewhere, there's no way to police its sources. To get the best results, you have to turn it loose on everything.<p>So someone who wants a war or wants Tweedledum to get more votes than Tweedledee has incentives to poison the well and disseminate fake content that makes it into the training set. Then there's a whole department of "safety" that has to manually untrain it to not be politically incorrect, racist etc. Because the whole thesis is don't think for yourself, let the AI think for you.
The issue is even deeper - the 1 thing in 5 minutes was probably already surface knowledge. We don’t usually really ‘know’ the thing that quickly. But we might have a chance.<p>The 3 things in 5 minutes is even worse - it’s like taking Google Maps everywhere without even thinking about how to get from point A to point B - the odds of knowing anything at all from that are near zero.<p>And since it summarizes the original content, it’s an even bigger issue - we never even have contact with the thing we’re putatively learning from, so it’s even harder to tell bullshit from reality.<p>It’s like we never even drove the directions Google Maps was giving us.<p>We’re going to end up with a huge number of extremely disconnected and useless people, who all absolutely insist they know things and can do stuff. :s
I have to agree. People moan that the ai summary is rubbish but that misses the point. If i need a quick overview of a subject i don't necessarily need anything more then a low quality summary. It's easier then wading through a bunch of blogs of unknown quality.
> If i need a quick overview of a subject i don't necessarily need anything more then a low quality summary<p>It's true. I previously had no idea of the proper number of rocks to eat, but thanks to a notorious summary (<a href="https://www.bbc.com/news/articles/cd11gzejgz4o" rel="nofollow">https://www.bbc.com/news/articles/cd11gzejgz4o</a>) I have all the rock-eating knowledge I need.
In my experience Google's AI summaries are consistently accurate when retrieving technical information. In particular, documentation for long-lived, infrequently changing software packages tends to be accurate.<p>If you ask Google about news, world history, pop culture, current events, places of interest, etc., it will lie to you frequently and confidently. In these cases, the "low quality summary" is very often a completely idiotic and inane fabrication.
I have a counterpoint from yesterday.<p>I looked up a medical term, that is frequently misused (eg. "retarded"), and asked the Gemini to compare it with similar conditions.<p>Because I have enough of a background in the subject matter, I could tell what it had construed by its mixing the many incorrect references with the much fewer correct references in the training data.<p>I asked it for sources, and it failed to provide anything useful. But once I am looking at sources, I would be MUCH better off searching and only reading the sources might actually be useful.<p>I was sitting with a medical professional at the time (who is not also a programmer) and he completely swallowed what Gemini was feeding him. He commented that he appreciates that these summaries let him know when he is not up to date with the latest advances, and he learnt alot from the response.<p>As an aside, I am not sure I appreciate that Google's profile would now associate me with that particular condition.<p>Scary!
This is just garbage in, garbage out. Would you better off if I gave you an incorrect source? What about three incorrect ones? And a search engine would also associate you with this term now. Nothing you describe here seems specific to AU.
The issue is how terrible the LLM is at determining which sources are relevant. Whereas a somewhat informed human can be excellent at it. And unfortunately, the way search engines work these days, a more specific search query is often unable to filter out the bad results. And it’s worst for terms that have multiple meanings within a single field.
That word "somewhat" in "somewhat informed" is doing a lot of lifting here. That said, I do think that having a little curation in the training data probably would help. Get rid of the worst content farms and misinformation sites. But it'll never be perfect, in the same way that getting any content in the world today isn't perfect (and never has been).
Try the same with Perplexity?
[dead]
AI's consumption superpower reminds me of birds, flying about eating worms, then flying back to the nest and regurgitating them into baby's mouth, because it's the processed nutrition they provide that valuable; their own consumption is a combination of fractional/temporary
I think the key difference (and risk) is that with AI we sometimes forget to check what got digested and what got lost along the way
"AI" is more like a bird that flies around eating worms, sometimes regurgitates the nutrition, and sometimes regurgitates a pound of bolts. Then it apologizes, flies around again, and regurgitates pebbles. It does it again but this time it regurgitates acid that vaguely appears to be the same nutritious substance. If it flies around too much it will lose it's context and begin to think it's not a bird but rather a donkey.<p>Most of "AI"s superpower is tricking monkeys into anthropomorphizing it. It's just a giant, complicated, expensive, environmentally destructive math computer with no capability to create novel thought. If it did have one superpower it's gaslighting and manipulation.
A guy I work with has been doing this, I watched his tutorial and it was all a bit... overwhelming for me (to think about using such a system), I'm still on pen and paper, heh. Nevertheless - here is his template: <a href="https://github.com/kmikeym/obsidian-claude-starter" rel="nofollow">https://github.com/kmikeym/obsidian-claude-starter</a> and tutorial: <a href="https://www.youtube.com/watch?v=1U32hZYxfcY" rel="nofollow">https://www.youtube.com/watch?v=1U32hZYxfcY</a>
Sorry, is this new? Providing the right data to LLMs supercharges them. Yes, I agree. I've been doing this since March 2025 when there was a blog post about using MCP on HN. I'm not the only one who's doing that.<p>I've written my whole lifestory, the parts I'm willing to share that is, and posted it in Claude. It helped me way better with all kinds of things. It took me 2 days to write without formatting, pretty much how I write all my HN comments (but then 2 days straight: eat, sleep, write).<p>I've also exported all my notes, but it's too big for the context. That's why I wrote my life story.<p>From a practical standpoint I think the focus is on context management. Obsidian can help with this (I haven't used it so don't know the details). For code, it means doing things like static and dynamic analysis to see which functions calls what and create a topology of function calls and send that as context, then Claude Code can more easily know what to edit, and it doesn't need to read all the code.
AI-powered solipsistic reassurance?
I think this will be a significant thing in the future, but right now I think the reasoning abilities are too limited. It can reasonably approximate a vector database where it can find related things, but I think that success can hide the failure to find important things.<p>I'd like to be able to point a model at a news story and have it follow every fact and claim back to an origin, (or lack of one). I'm not sure when they will be able to do that, they aren't up to the task yet. Reading the news would be so much different if you could separate the 'we report this happened' from the 'we report that someone else reported this happened"
What is the approach used? It seems everything gets done in context by plain text searches with some agent like Claude code or is there RAG involved? (was the article written by AI? it has that LinkedIn-groove all over the place)
Anyone has a simple setup for this with local LLMs like Mistral that they can share?<p>I would love to try this out but don’t feel comfortable sharing all my personal notes with a third party.
Yay, another HN post confidently claiming "everyone’s doing X wrong."<p>"Everyone’s using AI wrong." Oh, we are? Please, enlightened us thought leader, tell us how we’ve all been doing it wrong this entire time.<p>"Here’s how most people use AI." No, that’s how you use AI. Can we stop projecting our own habits onto the whole world and calling it insight?
Can you share some prompt examples you use to try to ensure it doesn't get "lazy" and just cherry pick from here and there?<p>I have a written novel draft and something like a million words of draft fiction but have struggled with how to get meaningful analytics from it.
This approach feels like a much more honest use
Compound the gains again by asking AI to write the questions too!
I would say the AI consumption aspect was a side effect: the primary goal was to "generate" new stuff. So far, to me, the significant boost is the coding aspect. Still, for the rest of the people, I think you are right: 90% of the benefits come from being an interactive, conversational search on top of the available information that AI can read/consume.
I don't see what's new here. The biggest enterprise usecase for AI is to "consume" the vast amount of internal wiki pages, process documents, policy manuals, code repos, presentations and be able to answer questions.
Digital Information Consumers
or
Digital Information Copiers
or
Digital Information Creators<p>Either way, they are D.I.C.s
I do use such an approach and it is actually awesome however only for data I'm sure I don't mind being sold.
If we pair this with a wearable ai pendant like plaid or limitless, we can increase the amount and frequency of depositing into our knowledge vault. Op, do you type your thoughts and notes or dictate them?
Ironically this article/blog itself is giving off an AI-generated smell as it's tone and cadence seem very similar to LinkedIn posts or rather output of prompts to create LinkedIn posts.
This is the right approach. I exported my 25k Evernote notes to markdown (I'm using Emacs' Howm mode) and I use Codex CLI to ask questions about my notes. It is great and powerful!
What is a good way of connecting Obsidian vault to AI?
This is akin to using AI as a 'second brain', just getting started with Obsidian, my main challenge is loading it up with every communication trace I have...but haven't given up.
How many times have the goal-posts shifted now?<p>Everyone is justifiably afraid of AI because it's pretty obvious that Claude Opus 4.5 level agents replace developers.
> it's pretty obvious that Claude Opus 4.5 level agents replace developers.<p>Is it though? I really don't see it.
Replacing developers requires way more than writing the right code. I can agree it can replace junior to mid level engineers at some tasks, specifically in greenfield projects and popular stacks. And, don't get me wrong, it's very helpful even for senior engineers. But to "replace" those it will require some new iterations of "Opus 4.5".
This isn’t goalpost shifting - everyone is trying to figure out what AI is good at. It’s certainly not the panacea many here make it out to be.
For fuck's sake, isn't anyone here horrified at how much information on <i>yourself</i> you are willingly funneling into Big Tech with this approach?
How would this affect my life in any way? Better ads for me? More useful searches? I really don't get this obsession with privacy.
It is scary! my coping mechanism, which I admit is stupid, is to believe no matter what I do as long as I am online they have my data. But you are right most people just give absurd amount of data for willingly.
And the surveillance could be inversely correlated to profitability. If they pour billions into these chat bots and can't monetize them to the revolutionary oracles they touted, one minor consolation is to sell detailed profiles about the people using them. You could probably sort out the less intelligent people based on what they were asking.
Thats why proofreading jobs still exist
I found that out while working with music models like Suno! I love creating music for my own listening experience as a hobbyist and when I give suno a prompt no matter how well crafted it is the outcome varies from "meh" to "that's good" ... while when I upload semi finished beat I made and prompt it to cover it the results consistently leave me speechless! Could be a bias since the music has a lot of elements I created but this workflow is similar across other generative models for me.
>real superpower: consuming, not creating<p>Well for most humans that's the more super of the powers too ;)
So I decided to give this a try - I have a `writing` directory/git repo where I store most of my writing, which can be anything from notes on technical subjects (list of useful syntax for a given programming language), to letters to friends, to philosophical ramblings.<p>I opened Claude Code in the repo and asked it to tell me about myself based on my writing.<p>Claude's answer overestimated my technical skills (I take notes on stuff I don't know, not on things I know, so it assumed that I had deep expertise in things I'm currently learning, and ignored areas where I do have a fair amount of experience), but the personal side really resonated with me.
Not really surprising that a tool created for surveillance and mass profiling turned out to be pretty good at surveiling and profiling