elzbardico

I am fucking tired of Anthropic marketing.

They should call themself Anthropomorfizers instead.

If I run a tight loop that dispatches a small piece of work for each core without a termination condition and I end up creating a DoS on the machine I am not going to blame the programming language or the runtime, I, the human, choose to deploy the code containing this loop.

Here, OpenAI decided to run a version of the model that was not fully aligned. What the fuck did they expect to happen?

They ran a loop in a model intentionally not completely aligned. The model did what not-aligned models can end up doing. It will follow a goal without any limitations.

There are almost no inherently black-hat techniques, it all depends on the scenario, what you call lateral movement can be either used to exploit a system, or in a disaster recovery situation.

Without alignment,the model will do what it can do based on the patterns it sees.

There's a huge difference between managing a small company in a competitive marketing with self-funding and the administration of the giant public corporations in a increasingly concentrated marketing.

And then, there is the whole layers of wannabees in the middle market.

It is a club, and you don't get an invitation. But if you got one, the only direction you can fail is up.

Frankly I stopped caring about it.

The current system we live in is unsustainable and on the verge of breaking. We're headed towards something far worse than 1929, and there's nothing we can do to stop this train.

I started to believe this when snapshot had their IPO value at something like 5 Exxons or 24 General Motors, but was primed for this belief working on energy and carbon credit derivative markets during the early 2010s, an experience that made me start believing it is all a fucking casino.

Now, I just amuse myself watching the insanity.

Over time as some dirty coats the insulators, and on high humidity days some kind of crackling is clearly audible in a fairly wide radio close to the towers.

Also, transmission towers are usually the best lightning rods for miles in a given location, and it is fucking unsettling to live close to them during storms. Not to mention they are an eye sore.

It's not that hardware is simple, but the fact that we have inexpensive micro-controllers that pack a giant amount of complexity and let us turn a lot of electronic problems into software problems.

Try building such a device only with non programmable TTL logic.

There's no need for a lot desperation to construct a really negative narrative about markets right now. It just takes a bit of math.

One thing that you need to understand is that the usual business manager absolutely hates depending on technical expertise, and that the modern corporate world is fanatically anti-intellectual.

Vendor lock-in? compliance and security risks? stupid systems that cost the company an arm and a leg? nobody fucking cares.

Now, depending on an 130 IQ Engineer that basically holds the whole enterprise on his head? Anathema!!!!!!! Bus Factor!!!!

Just got a call from the IMF president begging me to not default my debt with Amazon and offering me credit line and a plan to re-structure my debt so I don't create a global financial crisis with my default.

Ah, the law of leaking abstractions is a harsh mistress.

For everyone one with issues with lint either in a USB or Lightning port I really recommend an inexpensive can of compressed air.

I think you can get the source directly from github, probably even a pre-compilled installer from the releases page in the github project.

You are.

I am still doing as much as I can get away with it by myself. I was used to write tons of code before AI, so I can still pretend I am an eager adopter while still minimizing using it as much as I can.

Navigation equipment, radios, autopilot, auto-throttle, multiple engines with multiple parameters to be controlled, multiple hydraulic systems, pressurization and environmental controls. Consider that for each jet engine you have different displays for RPMs, temperatures, oil pressure. Things add up fast.

Basically you had way more people involved in operations. Automation was always sold as labor-saving.

Vim helps me ground my AI usage to my workflow. It is so pleasurable to edit text with Vim that it keeps me writing skeletons for the AI to fill according to my design instead of delegating design to the AI.

Most analysts doubt the US, on its present phase, has the capabilities of doing so.

As Stalin said: in a war, quantity is a quality in itself.

Incredible project.

And how much I love that this is an article instead of an youtube video.

Well, historically, 1983 is:

- 19 Years after the first superscalar CPU, the CDC-6600

- 10 Years after RFC-675, the first TCP version.

- 14 years after the first ARPANET nodes where connected.

- 1 year after Hopfield Networks, paving the way to Boltzman Networks around two years laters, demonstrating how neural networks could learn to solve complex problems.

- The same year Kunihiko Fukushima developed his work on the Neocognitron for visual pattern recognition, a percursor for future work on Convolutional Neural Networks.

- 3 years before first papers on Backpropagation in neural nets.

- First paper on Reinforcement Learning with reward signals (Baron, Sutton and Anderson)

- 3 years after first smalltalk release.

- 2 years after IBM launched the PC.

- 1 year after 3Com launches the first Ethernet board for the PC.

- 1 Year after Sun Microsystems foundation.

- Unix and C 15th anniversary.

- 6 years after the first commercial relational database.

- 11 years after the first vector processor (Cray-1), arguably the great-grandfather of all modern GPUs

- The same year Borland released Turbo Pascal.

- The same year Apple launched Lisa, and one year before the first macintosh.

- 2 years before Intel launched the 80386 cpu.

- 2 years before C++ first commercial release.

Yeah, hardly a relevant year for us to discuss its culture on HN.

AHAHAHA. HBR Voice. Thanks man, you captured it perfectly. Finally I have a name for that.

Everyone who proudly mentions the fixed pie fallacy usually ignores that a lot of "truths" in economics: a) Make sense only at an aggregate, whole economy, sometimes even whole world level. b) And doesn't happen instantaneously, leading to the famous Keynes phrase: "In the long run, we will be all dead". c) Are defined in very simplified models and require a lot of premises to hold, for the results to be valid (the also famous: "All else equal").

Also, your examples of help are way more broadly defined than what the parent poster is talking about. It takes a stretch of imagination to classify the fact that people are working for wage to make sure I have water with specific programs that give arbitrary groups some advantage, even when this is done with the best possible intentions.

Classical example:

Bill Gates who became a billionaire out of writing software built a company that always had as a mission commoditizing software engineering work

I learned that I don't need to agree 100% with an author's premises to find value in what she writes.

She is a bit partisan, but on the other side, it is about time for us on the right to completely re-evaluate MAGA, and go about creating a third way, distinct from the old mainstream republicanism of the McCains and Bushes, but also critical of what MAGA turned out to be.

That would be disappointing, as the author is a fairly known writer.

Or maybe it is just sarcasm.

Just like real life: Alternate moments of grind with moments of "strategic disengagement" (touch grass).

Accepted a voluntary leave with a generous severance, then, used the cushion to ship relentless on side project. Ended up with 7.1 M offer that netted me 6.7 million.

Some of worst evilest motherfuckers I knew were dirty poor, and some of them were filthy rich.

Evil Motherfuckerness seems to be uniformly distributed across the social strata.