Last week-end I was exploring the current possibilities of automated Ghidra analysis with Codex. My first attempt derailed quickly, but after giving it the pyghidra documentation, it reliably wrote Python scripts that would alter data types etc. exactly how I wanted, but based on fixed rules.
My next goal would be to incorporate LLM decisions into the process, e.g. let the LLM come up with a guess at a meaningful function name to make it easier to read, stuff like that. I made a skill for this functionality and let Codex plough through in agentic mode. I stopped it after a while as I was not sure what it was doing, and I didn't have more time to work on it since. I would need to do some sanity checks on the ones it has already renamed.
Would be curious what workflows others have already devised? Is MCP the way to go?
Is there a place where people discuss these things?
I've been wondering the same. Apparently, where I live (HCOL) people don't care about ever paying off the full amount, as long as the monthly installment is low. Regulations only require that you paid off 33% of the house value after 15 years.
You do need to start at 20% capitalization though, which at current prices means they need to own assets already (market risk), or have bonkers amounts of currency sitting around (inflation risk). I understand the FOMO some people must have.
I wonder if ultimately there is a cascade happening, where increased valuations lead previous owners to either take new mortgages, or sell their old house, and pay more for a new one. Which creates a cycle of ever increasing prices? I've never read anything though whether such an effect exists.
In HCOL areas of the US, 10% down is becoming more common with private lenders and 5% is even a possibility (or at least was pre-pandemic). Pre-pandemic we had multiple competing banks for a $1MM+ mortgage w/ 10% down.
When I read stories like this about cool technology from the past that people can still enjoy today, I think it's a pity that the source code is locked away and lost as people and companies move on to new things.
IMO, there should be some kind of archive that conserves and publishes them after some time has passed, so that they could be ported to new hardware and kept accessible. and somehow documented for future historians.
Re-releasing old games on new systems is pretty common. Crash Bandicoot got re-released on the PS4 for example. So companies are going to hold onto their assets in hopes of being able to cash in on them again.
Not all of them even have it. Once everyone moves on from a project unless the company has an archivist that stuff tends to get thrown away or forgotten about in some dusty drawer in some box of DAT tapes that everyone forgot about. I would say it would be around the very late 90s that companies started realizing they should hold onto that stuff. Even then...
> What neuromorphic hardware can I buy to run your code/ the SNN?
Current neuromorphic hardware is not easily accesible, but you can simulate spiking neural networks. Check out, e.g. https://brian2.readthedocs.io/en/stable/ or Nengo.ai
Empirically, you could have bought a share of the SPX at any point in time, and sold it with profit later. The real problem is what happened in the time between, and whether you were able to hold on.
Nice reminder of some general principles. I wonder how well it applies to young scientists given today's academic career prospects, though.
> If you want to make progress in any area, you need to be willing to give up your best ideas from time to time. [...] Medawar notes that he twice spent two whole years trying to corroborate groundless hypotheses.
Unfortunately, being willing to scrap your idea is only one part of the equation. Securing funding after 2 "failed" post-docs is an entirely different matter.
This is why you should not ever dedicate yourself to a single project no matter how much you believe in it. You should at a minimum have one back up project that will generate publishable results no matter what happens.
Ouch. Your point is tough to stomach, but there is another side to it they don't teach you in school: tenure-track research isn't the only path. In fact, tenure-track research may actually be the worst path for the vast, vast majority of PhD graduates.
Exactly. It's foolish to follow this path. There's a ton of VC money out there, if you need millions or billions to see your idea through. If not just make your own money. Honestly feels less stressful and hypocritical than the academic Ponzi scheme.
Few academic disciplines lend themselves to monetarisation via startups. Indeed, many academic disciplines don't lend themselves to any kind of monetarisation, except academic funding.
However, building a business is a different problem from doing research. When one accepts VC funding, the funding comes with the expectation that it will lead to a high-growth business. This is fine when you have an idea that has the potential to “take off” from a business standpoint.
However, there are many research problems where there is no obvious or immediate business application. The aim of such research is different from the aim of investors. This requires a different funding source, one that is willing to embrace the risks that come with research and is willing to do work solely for the advancement of science, with productization being a nice side effect rather than an expectation.
Of course, obtaining such funding is not easy. Part of what makes modern academia such a rat race is because of how competitive it is to procure research funding from funding agencies such as the NSF (disclaimer: this is a US point of view; I’m not very familiar with the situation abroad). My advisor works hard applying for grants, and sometimes they get rejected. I’d love a Genius Grant ($125,000 a year for five years with no strings attached) to work on whatever research I want without any pressures from the funding agency or from managers, but there are only so few awarded per year.
There is also the matter of research freedom in the sense of being free from the pressures of short term thinking and "publish or perish" mentality. I am reminded of Alan Kay's observations (http://worrydream.com/2017-12-30-alan/) about short-term research. I'm also reminded of what the discoverer of the electron, J.J. Thompson, once said in a 1916 speech that resonates with me whenever I think about research:
"If you pay a man a salary for doing research, he and you will want to have something to point to at the end of the year to show that the money has not been wasted. In promising work of the highest class, however, results do not come in this regular fashion, in fact years may pass without any tangible result being obtained, and the position of the paid worker would be very embarrassing and he would naturally take to work on a lower, or at any rate a different plane where he could be sure of getting year by year tangible results which would justify his salary. The position is this: You want one kind of research, but, if you pay a man to do it, it will drive him to research of a different kind. The only thing to do is to pay him for doing something else and give him enough leisure to do research for the love of it."
For me, my dream is to start and grow a non-research lifestyle business that pays the bills, so that way I can spend the rest of my time on research, which is what I'm passionate about.
I am in the same path, make the money and fund my own research. All other models have been perverted by existing long enough that the culture around them is very antithetic to the true motivation of the endeavor.
The only systems not amenable to this model is perhaps high energy physics or things of that sort but I'm sure some resourceful mind will come up with ideas!
Exactly. Most academics really can't win. It's how the system is designed.
This is why I sold out after my MD/PhD and became just a regular physician. Maybe in a few decades I'll have enough in savings to fulfill my dream of becoming a mad scientist...
Last week-end I was exploring the current possibilities of automated Ghidra analysis with Codex. My first attempt derailed quickly, but after giving it the pyghidra documentation, it reliably wrote Python scripts that would alter data types etc. exactly how I wanted, but based on fixed rules.
My next goal would be to incorporate LLM decisions into the process, e.g. let the LLM come up with a guess at a meaningful function name to make it easier to read, stuff like that. I made a skill for this functionality and let Codex plough through in agentic mode. I stopped it after a while as I was not sure what it was doing, and I didn't have more time to work on it since. I would need to do some sanity checks on the ones it has already renamed.
Would be curious what workflows others have already devised? Is MCP the way to go?
Is there a place where people discuss these things?