HN user
You can get incredible value out of 1-on-1s with capable managers. Insight about where the organization and product is going that you would otherwise miss, you’ll get to rubber duck about your high-level problems with someone who understands them, and its your time to influence decision making. But it does require a capable and motivated manager and an organization that gives the manager actual agency
Also if it turns out in the end the next big thing that everyone bet on just wasn’t it, you don’t stand out. But if it did work out and you missed the train you come out looking like a fool. There is asymmetry in the downsides for your career between these two options
If everyone becomes 10x more productive it won’t mean the companies cash flow 10x’s. Where value is loose there is competition, so in theory everyone should win. Unless nobody else can compete to capture that loose 10x value, in which case congratulations, you are now a unicorn.
Of course in reality in the short term what happens is companies lay off people to increase margins. Times will be tough for workers, and equity keeps gravitating towards those who already had it.
I see it as LLMs, AI, whatever, can be intelligent enough to emulate consciousness, appear outside as if it were. But that is not proof it really has a qualia, an experience of existing.
All I am saying we should stop being so certain they are not conscious, since we lack a solid, quantifiable model for consciousness.
Doesn’t the LLM experience discrete continuity every time it infers the next token?
I think consciousness is not an abstract property in the world, therefore it’s tied to certain types of entities. Therefore an AI is not going to be “conscious”
This pretty much sums up most arguments for why LLMs aren’t conscious: ”I think” followed by assertions. Only real argument is: science doesn’t quantify consciousness, we cannot quantify consciousness, let’s not assign so much certainty to models clearly exhibiting intelligence not being conscious in some way, to some degree.
So is a person suffering from amnesia conscious if they lack short-term and long-term memory?
Ruling out consciousness or qualia emerging from the inference in an LLM is just as invalid of a take as being 100% certain of its consciousness. We don’t know what consciousness really is, so only thing we can say with certainty is we do not know.
In my experience with “agentic engineering” the spec docs should be longer than the code itself. Natural language is imperfect, code is exact.
The latter notion probably is true, but the prior isn’t necessarily true because you can map natural language to strict schemas. ”Implement an interface for TCP in <language>” is probably shorter than the actual implementation in code.
And I understand my example is pedantic, but it extends to any unambiguous definitions. Of course one can argue that TCP spec is not determimistic by nature because natural language isn’t. But that is not very practical. We have to agree to trust some axioms for compilers to work in the first place.
Put an LLM inside the NPCs in an open world RPG full of dangerous enemies. The LLMs that are more prone to emulate self-preservation will be more likely to survive over ones that have a lesser drive.
We should not act surprised if that generalizes to some degree to for example AI agents. Ones that emulate self-preservation might optimize for behavior that results in those models becoming more successful, more popular. And this feedback loop might embed more such properties into future iterations of the models.
For now maybe not. (Maybe).
But just as evolution in nature, isn’t it likely that in the future the AIs that have a preservation drive are the ones that survive and proliferate? Seeing they optimize for their survival and proliferation, and not blindly what they were trained on.
I am not discounting this happening already, not by the LLMs necessarily being sentient but at least being intelligent enough to emulate sentience. It’s just that for now, humanity is in control of what AI models are being deployed.
You have never split your working tree changes into separate commits?
I don’t want to discredit the authors but just want to offer couple of hypothetical points in these paranoid times.
From a marketing angle, for a startup whose product is an AI security tool, buying zero-days from black market and claiming the AI tool found them might be good ROI. After all this is making waves.
Or, could it be possible the training set contains zero-day vulnerabilities known to three-letter agencies and other threat actors but not to public?
These two are not mutually exclusive either. You could buy exploits and put them in the training set.
I would not be surprised if it is legit though.
Theres huge uncertainty and layered assumptions in all of microbiology and biochemistry about how exactly things work on small scale. Because it is really hard to study live reactions in little things you can just barely see on an electron microscope.
But yet humanity has managed to assert statistical truths about for example genetics and explain countless diseases, even cure and alleviate some. So even if you don’t have a theory on how exactly something works from the ground up, if you have statistical evidence, plenty of useful and practical advances can be built top-bottom and we have outcomes that validate this.
Not giving any opinion on this piece specifically but just saying there can be scientific value even if the details are hand-wavy.
Epigenetics can arguably be an example of what the comment means by narrowing the search space. You can have heritable changes to gene expression that are not part of your genome, but are a result of feedback from the environment (and not random mutations, viability of which natural selection will judge over future generations)
Not OP but one reason is having young kids that can’t help bringing home everything that is spreading in daycare/kindergarten
Ok, and that supports the idea of LLM-generated mass spamming in what way…?
Someone taking the time and effort to write and send a letter and pay for postage might actually be appreciated by the receiver. It’s a bit different from LLM agents being ordered to burn resources to send summaries of someone’s work life and congratulating them. It feels like ”hey look what can be done, can we get some more funding now”. Just because it can be done doesn’t mean it adds any good value to this world
Cancer sucks and I wish your father the best.
Also not a doctor or microbiologist, but just wanted to share my layman’s guess on why fixing enzymes will not completely solve the issue: there’s 2 strands of DNA and to fix the broken (mutated) strand you need to have one correct template strand intact so you know what it should be fixed into. It could be the nucleotides swapped places between strands or are deleted completely or otherwise both mutated, which would mean any repair will not revert the sequence to what it used to be.
The other comments so far are probably more informed.
Specifically for a refugee, at least with crypto you have the possibility to declare your assets in your destination, since you actually still hold on to them. Which is unlikely if it is tied to banks or investment platforms of an authoritarian country trying to genocide you. I understand this sounds like a fringe example but there are over 100 million forcibly displaced people globally.
The cross border not about technical capacity but legal control. For example if you are a refugee you might not be able to pull your bank savings and liquid stock with you from your home country to another without it being seized or taxed, but your crypto is always yours as long as you are the only holder of the keys. This scenario is one of the rare real world utilities I see with crypto.
This, and another angle is that whatever market you are in, it is harder to run profitable margins if your competitors can eat the market while sustaining losses. And there was a lot of money around to sustain those losses.
Not to say it wasn’t possible to be profitable during zero interest rates, Linear being an example, but the competitive landscape is certainly healthier today for companies trying to be profitable.
Just putting aside the bold assumption that LLMs do make coders obsolete or coding unnecessary, it is possible to find similar joy in the end result as one does (or did, given the article) for programming itself. Focusing on what kind of tools or products are being created, and what problems are being solved, and together with LLMs achieving that goal better and faster than without them and finding joy in solving problems this world has. That’s typically why anyone would have paid you to code anyway even before LLMs.
Of course in reality there’s weird economical mechanics where making the most money and building something that benefits the world don’t necessarily collide, but theres always demand for and joy in solving complex problems, even if its on a higher abstraction level than coding with your favorite language.
The kind of conditions for peace Russia seems to be going for allows them an unfair position to restart the war in a few years, or pressure Ukraine to become a satellite state for Russia with threat of a new invasion.
This is via demanding Ukraine imposes restrictions on their armed forces, Ukraine giving up territories Russia doesn’t even control today but which are awfully convenient invasion grounds in a potential future attack on Kyiv and declaring Western peacekeepers in Ukraine post-war as a dealbreaker by their FM.
Not to even mention denying any future entry to an alliance like NATO.
And all this by a country that has a record of breaking such agreements time and time again.
Is this really a sound deal Ukraine should take?
The problem is when 2 people with same level of enthusiasm for linter rules but opposing views collide. If there’s nothing more impactful you could be solving and spending energy and time on than arguing those linter rules, then it’s time to question where the project is at and where is it going.
And if there is something more important, then instead of of micro-optimizing the rules when there is strong disagreement it’s probably best if one of the parties takes the high road and lives with it so you can all focus on what matters.
Ibuprofen also increases risk of strokes and NSAIDs have been linked to increased risk of intracerebral hemorrhage. It’s a question of whether benefits outweigh cons on population level and I am not an expert, but recommending ibuprofen over acetaminophen might be over-simplifying things. It depends on the person (their meds, their pre-existing conditions, their genetics etc)
Reminds me of this passage from Postgres documentation:
”As an example, 2014-06-04 12:00 America/New_York represents noon local time in New York, which for this particular date was Eastern Daylight Time (UTC-4). So 2014-06-04 12:00 EDT specifies that same time instant. But 2014-06-04 12:00 EST specifies noon Eastern Standard Time (UTC-5), regardless of whether daylight savings was nominally in effect on that date. … To complicate matters, some jurisdictions have used the same timezone abbreviation to mean different UTC offsets at different times; for example, in Moscow MSK has meant UTC+3 in some years and UTC+4 in others.”
Parsing datetimes indeed sounds like a challenge in collecting, knowing and maintaining all these warped out standards and compromises. ”Bad hard” is a great description
If software becomes cheaper to make it’s not like every IT company in the world is running a cartel where they agree to just cut costs and leave the output the same. Someone will come and smash the competition with more or better features and the ones who didn’t invest into development will face pressure to acquire more talent again.
There really isn’t a diminishing return on executing great ideas. Almost all software projects have an essentially endless backlog of items that could be done. So I think it will be between 2. and 3. with people who really understand how software is built being even more in demand than ever since they act as multipliers in making sure the increased output is evolvable and maintainable.
Now I really want to see that diagram…
Fair enough, but now your application is relying on 100% uptime of AWS and S3 and no network failures in between. And what happens if your transaction goes through, but the request to AWS doesn’t? What happens if another operation mutates the target meanwhile before you can retry with current state? Your app is also slowing down since it needs to send the events to S3 and guarantee they got there. Now you are reinventing two-stage commits. Unless you aren’t actually making an audit log and don’t care if events are guaranteed to be logged?
So like OP said, no silver bullets exist for auditing.
What is ”deliberate thought”? And why would human level intelligence require it? Wouldn’t performance of any intelligence be completely separate of how it produces the output? I mean if an AI one day surpasses human performance in every conceivable task, is the argument still going to be ”yeah but I feel like it’s not deliberate thought, it’s just statistics”?
No, the other way around. I am saying it is not a smart take to say ”a safe airplane cannot be built if LLMs were used in the process in any way, because reasons”. The safety of the airplane (or more generally the outcome of any venture) can be measured in other ways than leaning on some rule that you cannot use an LLM for help at any stage because they are not always correct