xlbuttplug2

I'm guessing the next generation of US frontier models will be heavily anti-distillation at the cost of user experience (significant rate limiting, more flat out refusals, hidden thinking, etc).

And as long as they maintain a significant advantage in capability, we will continue to kiss the ring.

May not be very effective if so.

I'm assuming finding vulnerabilities in open source projects is the hard part and what you need the frontier models for. Writing an exploit given a vulnerability can probably be delegated to less scrupulous models.

It's also trading at a huge premium. Probably worth a read if you're considering it: https://www.morningstar.com/funds/fundrise-innovation-is-not...

My hunch is it's a combination of

* coming in with a bias of not wanting it to work

* having too high of an expectation

* giving up too early

* not trying SOTA models

* not taking the effort to communicate intuitive or painfully obvious things

But perhaps it is too dumb to solve the type of problems you guys are working on and no amount of cajoling will help. All I know is "it works for me."

In my case, the volume of code is roughly the same. I'm not using the efficiency towards pumping out more code, just using it to be AFK more.

I spend enough time iterating and refining to the point I'm comfortable taking ownership of the outputted code. Perhaps hypocritically, I do mald when people upload code for review that they clearly haven't taken the effort to read through critically.

More like a day's worth of coding condensed into half an hour. Time to review/test is mostly unchanged.

Day to day is mainly minor feature additions into a stable product so not a huge amount of code churn.

Is that you, boss?

People with a lower multiplier are either in the minority of developers solving genuinely hard/novel problems or, more likely, they've just not figured out how to tap into AI's potential.

Granted, to your point, a decent chunk of the HN crowd belongs to the former and can't relate to us paycheck stealers.

they aren't even making 1.5x developers

I won't try to speak for anyone other than myself, but my multiplier is definitely over 1.5x, probably higher than 5x.

I choose to sit on my hands in my freed up time so upper management does not catch on to and exploit this fact. Eventually they will though via overzealous coworkers.

My probably incorrect, uninformed hunch is that users convinced of how AI should act actually end up nerfing its capabilities with their prompts. Essentially dumbing it down to their level, losing out on the wisdom it's gained through training.

"But we anonymize prompts before training!"

Meanwhile the prompt: Crop this photo of my passport

Or simply embrace ignorance. Why hold on to things you don't use? Accept AI as an extension to your brain, and let the now dormant parts atrophy.

Yes, you will be vulnerable should you lose access to AI at some point, but the same goes for a limb. You will adapt.

Huh. I'd never thought of this. If that is actually meaningfully beneficial, I wonder if they'd design self driving cars with the seats facing backwards, given there's no longer a necessity to look at the road.

(edit: I guess it's more of no-brainer on a train/bus where you don't have a seat belt)

Would you have a problem if AI was no longer a tool but the artist itself (i.e. no human intervention)?

People seem to have an irrational fear of being entertained by AI, equating that to admitting that it is a higher form of intelligence than their own.

Surely they've reserved the best models for themselves and have people looking into how to optimally harness untapped potential from LLMs?

Edit: I guess the competition between them keeps them honest and forces them to release their best models so they don't lose face.

They can and most likely will release something that vaporises the thin moat you have built around their product.

As they should if they're doing most of the heavy lifting.

And it's not just LLM adjacent startups at risk. LLMs have enabled any random person with a claude code subscription to pole vault over your drying up moat over the course of a weekend.

You can still do all these things manually. Now you just have the option not to.

The article says the man was trying to see the squirrel's back implying a larger radius.

But yeah if your circuit completely fits inside the other person's circuit, then you've been gone around, no matter how slow or fast you both are.

That the man technically went around the squirrel without ever having caught up to it.

Say instead of just walking, the man was laying down a net/barricade around the tree. As soon as the man completes the circumference, the squirrel must admit that it has been gone around.

I have very fond memories of playing the Pokemon games, but I always saw the battling aspect as a hurdle to unlocking more of the story/world, which was the true appeal to me. I was content turning my brain off and overleveling my mons. Different strokes, I guess.

I have the exact opposite prediction. LLMs may end up writing most code, but humans will still want to review what's being written. We should instead be making life easier for humans with more verbose syntax since it's all the same to an LLM. Information dense code is fun to write but not so much to read.

Even lay-er person, but maybe the specificity is not that impressive in mice? Perhaps when you scale to more complex animals it is inevitable to see false positives (detrimental effects to healthy cells)?

I think it has to get good enough to a point where humans are not the bottleneck for code review and course correction.

I guess the "velocity" multiplier is closer to 10x rather than the 1000x needed for true disruption capability.

I think the turning point will be when AI assisted individuals or tiny companies are able to deliver comparable products/value as the goliaths.

Yes, and we'll continue to have human coding competitions for entertainment purpose. Good luck trying to live off the prize money though.

I wonder if someone's looked into the optimal time of day and day of the week to post for maximum traction.

If I had to guess it would be monday morning pacific time when people would rather be doing anything than working.

That's probably intentional to prevent you from realizing how unsanitary the room really is.

Even if the prompts are technically leaked to the provider, how would they be identified as something worth optimizing for out of the millions of other prompts received?

It does not prevent a bot from using your ID. But a) the repercussions for getting caught are much more tangible when you can't hide behind anonymity - you risk getting blanket banned from the internet and b) the scale is significantly reduced - how many people are willing to rent/sell their IDs, i.e., their right to access the internet?

Edit: ok I see the argument that the feedback mechanism could be difficult when all the website can report is "hey, you don't know me but this dude from request xyz you just authenticated fucked all my shit up". But at the end of the day, privacy preservation is an implementation detail I don't see governments guaranteeing.

I did think asymmetric cryptography but I assumed the validators would be third parties / individual websites and therefore connections could be made using your public key. But I guess having the government itself provide the authentication service makes more sense.

I wonder if they'd actually honor 1 instead of forcing recipients to be registered, as presumably they'd be interested in tracking user activity.