secnicholas

28
- Free Software and Security enthusiast
- Autodidact
- Building a FOSS homelab SOC/SIEM
- Ghost in the Shell/Motoko Kusanagi fan

[secnicholas @secnicholas_x]

my takeaway from this is that if you want to provide the most value with the least resistance you should ensure you pile up enough simplicity until it becomes indistinguishable from complexity and still entrances the idiots.

[Neet @neet_sol]

it’s insane how much effort you need to put in just to land some mid corporate job

[thaddeus e. grugq @thegrugq]

Security is maintained through the combination of

- people
- technology
- procedures

Fixation on bugs as the sole element of security (offensive and defensive) is a category error. A profound misunderstanding of how systems work, and how they fail.

[Lachlan Phillips exo/acc 👾 @bitcloud]

Why are we debating this. You can calculate the outputs of an LLM by hand.

It's not conscious. It's a math problem.

[Brad R. Torgersen @BradRTorgersen]

Your weekly reminder: AI is not using reasoning. Nor is AI moral. AI is just endless pattern-matching and mimicry. There is no mind at work in those data centers. Nothing even close to intelligence. It's just data churn. And the data churn will spew wrong answers at you.

[Web Design Museum @WebDesignMuseum]

Fight Club flash website in 1999

[Hash @BitBangingBytes]

SCANNING ELECTRONS in my GARAGE!

[BladeoftheSun @BladeoftheS]

For anybody who doesn't know AI is only free because they need you to use it.

Once they have everyone dependent on it, it will become really expensive really fast.

Just like Uber, Netflix, Groceries...

[Eric S. Raymond @esrtweet]

There's been a lot of debate about whether LLMs are conscious lately.

I'm not going to say that I know what consciousness is in an ontological sense, because I won't claim that until I can propose an experiment that reliably distinguishes conscious systems from unconscious systems.

I think most talk about "consciousness" is just gabble leaning on a bunch of ungrounded terms. Not even worth the time to listen to, much less debate.

However. There is one thing I think I do know from the phenomenology of being human. That is that consciousness in humans requires two things: (1) a running loop of evaluating sensory data and fitting it to models already in the brain, and (2) at least a limited ability to introspect on that loop. One might say: consciousness is the mind watching the mind adapt to the continuing challenge of reality.

Interrupt the loop, say with anesthetics, and consciousness departs. Resume the loop, consciousness resumes.

Human-like consciousness requires continuous sensory input, constant feeding of the brain's predictive engine. Human minds disintegrate if you put them in isolation tanks for too long.

Do these observations permit us to say anything about whether LLMs might be conscious?

I think it does. LLMs do not run a continuous predictive loop. They are stimulus-response state machines the activating response to a prompt, do computation, present the result, and then go quiescent waiting for the next prompt.

I think this disqualifies them from having anything that we can meaningfully analogize to consciousness in humans.

However, I definitely do not exclude the possibility that future artificial intelligences with different architectures and a different embedding in the physical universe will be conscious in a meaningful sense. In fact, I rather expect I will live to see it.

[signüll @signulll]

software used to have so much friction that building things selected for a very specific type of often strange person but ai removes most of that activation energy altogether. now the unit of creation is i have an idea, i ask for it, & i keep changing it.

& once someone experiences that you can feel their relationship with computers starting to change. e.g. a lot of my non technical friends are making things with claude that even surprise me often.

that behavior feels way more important than vibe coding itself which is no longer even vibe. it’s actual coding.

[Panat @ptaranat]

yeah we use linux at the bookstore how could u tell

[Perry E. Metzger @perrymetzger]

If you ask an LLM to describe its vacation with its family in Italy, it will do so. It will happily explain what sort of practices it found helpful when lecturing at a university, or how it drafted contracts in the past, or what herbs it likes the taste of in that dish you are cooking.

(Some LLM systems have been extensively RLed out of doing this, but the fact that you have to do that in the first place is significant; by nature, they happily confabulate.)

An LLM’s supposed description of its internal states should thus be viewed with extreme suspicion, as should the motivations and rationality of anyone who takes such descriptions seriously.

[secnicholas @secnicholas_x]

i struggle to fully agree with the take that "they reduce you're ability to learn/will weaken your cognitive skills" but those cons are more likely to be true if you do any of the above imo

[secnicholas @secnicholas_x]

imo if you're using LLMs to dev/configure, avoid all of these:
- using more than one agent unsparingly
- having an agent manage other agents
- using background agents
- letting it build before you've planned a robust skeleton/put your constraints into words

[secnicholas @secnicholas_x]

LLMs + Unix philosophy + open source is a heavenly combination

[secnicholas @secnicholas_x]

@GergelyOrosz imo it's only manageable/can get close to irrelevant if you're successfully using agents to speed up what used to be a time cost for increasing composability and adding testing as they're what shrink debt the most

[kache @yacineMTB]

The differentiating skill today, in the age of AI does your work, is the ability to understand. The ability to know where your understanding ends. A high standard for your understanding itself

[Casie Nakamura Inc. @kwargq]

This is a prelude to Empero's own graphical harness as an extension on top of our Abacus agent. Comfy UIs are for everyone

[Casie Nakamura Inc. @kwargq]

Shitty designs for shitty things

[secnicholas @secnicholas_x]

@kwargq this is so good!

[Abdulkadir | Cybersecurity @cyber_razz]

An adblocker you can slide into the back of your router.