- The memory shortage might be what kills CXL or at least makes adoption slow down.
- Cz furnace for crystal growth, or an open source curve tracer and semiconductor parameter analyzer.
But I think that right now the cost of materials is just to high, even outside of tech.
The cost of metals, cement, and power can be heart stopping for the younger folk.
- 30% of the manufacturing of consumer memory has shifted to enterprise and big tech market, so I mean this was basically a given.
Its a shame we can't make chips in the basement with affordable tools like you can if your a hobby machinist or woodworker.
- Does anyone know or have they leaked potential cost of acquisition?
- AGI is as possible as a quantum computer, a dyson sphere, alien life, and room temperature and pressure supercomputers.
Sure in theory this is possible....but people forget this tech is no more disruptive than any "paradime shift" like the wheel, the invention of the blast furnace, the industrial revolution.
New paridimes care not for politics...I would argue it is humans that are making this a political issue.
We need to step back and learn how to separate innovations and major advancements from politics. Politics has become a limit that is holding us back and dividing us.
- This is interesting in theory....but will probably help capture markets for those with money.
You can't have a new search or open source AI if you can't afford to pay scraping fee's.
- Putting mass into space is expensive.... And given how low gravity is on some places like the moon....wouldn't going to the moon first make more sense.
The moon is literally free mass, its already in space and we just need to refine it.
The moon is also closer to the earth so it can be used as a launch off point.
See the work anthrofuturism has done on YouTube.
- As someone who is studying security, part of me has a child like awe for the sheer complexity and magic that is semiconductors and chip manufacturing.
Thank you for sharing this, I will be giving it a full read.
- Thanks Steve!
- I get manufacturing small stuff is hard.
But you would think we would see more competition in trying to reduce the barrier to entry, by say looking at making this manufacturing possible for a maker space or homelab and not just government funded and subsidised factories, of which only a small number exist.
- I think the real bubble is NVIDIA....not AI.
Is AI overhyped...sure it is just like with any new technology like quantum computers or crypto or big data. The use cases are real and the value exists....its just because it is so new that everyone wants a slice of the pie and so we have the situation like we are in now.
AI, just like big data or crypto...is very hardware expensive. While we have gotten better at inference and training, the real issue is the cost of chips. These hardware people know what they have is in high demand and so they are charging through the roof.
- Agree but I would argue RISC is catching up fast.
- Awesome tool and team you have.
Few questions I though of, and I apologize if they seem stupid as ML is not my focus of study.
1.Has your team ever considered formal verification of code to show how reliable the process you have is?
2. If data has been removed via your pipeline, is it possible to still infer the type of data based on position or format? (Names of people being located in certain places of a sentence, or say the fact the data is formated a certain way could reveal its a date or timestamp?)
3. You mentioned clients can deploy via VPS, does that mean this is a fedramp ready product? (Do you see this tool being offered to public institutions?)
4. Do you have any internship openings for college students in the summer of 2026?
- Given how arm license is know to be less than friendly.... Wouldn't it be preferable to explore a RISCV architecture.
As far as I know RISC provides similar power efficiency and sleep that is like ARM.
- 3 major problems no matter what you do that can't be avoided.
1. Water will rust metal. Some allows resist corrosion but they cost more which increases cost and engineering complexity.
2. Water freezes at a higher temp than diesel.... Which means you need to worry about keeping the water warm enough it dosnt freeze or turn into a slushie.
3. Water varies based on location. I doubt this works for salt water but let's say we decided to add water to the combustion process....now you need to worry about hard vs soft water and how that wears down everything or impacts efficiency. Or do you need reverse osmosis or distilled water?
- Do you know what algorithm or sensors it uses to calculate BP?
- I'm taking upper level classes, and had an issue where the professor asked me to recall some information from an old math class. I had to be honest with him, so I told him straight up in front of the class: "I took the class and passed it, my passing of a class is not mean I actually remember or understand what was taught. While I do have an interest in math, I struggle with it and forget what I don't regularly practice. I can't afford to constantly retake classes, and as such I have sacrificed my ability to know X subject in math in exchange for the passing grade I got"
I can't be regularly be practicing all levels of math and reviewing on top of my regular class schedules and work....I literally don't have the time, mental energy, or ability to even remember any of this stuff.
- Issue is, unless you can be 100% sure you hardware has not been built with a vulnerability or backdoor, or subject to an evil maid attack....then you can't be sure its trustworthy.
- I sort of agree. I am using gadget bridge with my old watch and while it works fine in general...the step tracker,and calories burned seem really inconsistent.
The heart rate and oxygen saturation seem accurate compared to other devices measuring it.
I wouldn't mind a new watch that was more expensive than my current one if it was more accurate and had better setup compared to how I had to pair my current device. I'm not talking google watch or apple watch, but sub $300 device.
A lot of silicon on a GPU is dedicated to upscaling and matrix multiply.
Ultimately GPU's main use is multimedia and graphics focused.
See all the miners that used to do GPU based mining...or the other niche markets where eventually the cost of custom asic becomes to attractive to ignore even if you as a consume have to handle a few years of growing pains.