Avoid purchasing hardware for ML and AI
I had this conversation with my cousin who wanted to purchase a high-end laptop with the latest processor and graphics card. He already had a Mac book but he couldn't get some ML stuff working on that. His idea was to purchase a brand-new laptop that had all the latest specs and he did not worry about the cost. Unfortunately, we did not have enough options in India. We were thinking of importing laptops from other countries where it was already released. We have been continuously hunting this for almost 2 months. Later we had a conversation if it would be worth the effort or not.
The idea was to purchase the laptop, improve his ML skills, and then switch to a different job. He is well settled in a good job but he wants to upskill and was looking for upgraded hardware here. Ultimately after gathering some suggestions and consulting some experts, I asked him to stop investing in hardware and instead use cloud solutions. I understand that it may not be easy to get everything on the cloud but these days every other cloud solution is providing a competitive option. First, let's see what and all would be the problem if we invest in hardware.
Hardware can get outdated
This is the biggest problem. Today we might think that we have bought the latest technology and the best specs available in the market but there is a high chance that it might soon get outdated and we end up staying outdated. But now it would be so tempting to purchase another laptop or other hardware that is recent. If this was the thought before 10 or 15 years then it is natural because we did not have enough options but today we have cloud solutions and we get almost everything for rental through a cloud service provider. Investing in heavy-priced hardware can be avoided.
Depending on cloud solutions
There are also situations where we may not need the hardware we purchase for a longer period. We may not even be able to spend enough time on our investment if we already have a day job. In that case, having a cloud solution like a server on AWS and doing our stuff there would be the best way to do it because we can turn the service on and off whenever we require it. There are also competitive options available on cloud service providers like AWS. The support system is also very good and if we need our hardware replaced, we can easily do it if we are renting it. For physical hardware, it may not be the case and if there is a fault, we have to either live with it if possible or take a lot of effort to replace it.
Technology keeps changing
As I mentioned earlier, technology keeps changing. If we purchase something and try to be happy, there will be another bigger or better thing emerging in the market tempting us to buy. Obviously, we cannot afford all the technological changes that are happening around us and it is also hard to stay up to date. We can keep chasing technology only to some extent. Especially ML and AI are very new and they will keep changing rapidly in the next decade.
If you like what I'm doing on Hive, you can vote me as a witness with the links below.
 |
 |
 |
 |
 |
 |
Posted Using InLeo Alpha
This post has been manually curated by @steemflow from Indiaunited community. Join us on our Discord Server.
Do you know that you can earn a passive income by delegating to @indiaunited. We share more than 100 % of the curation rewards with the delegators in the form of IUC tokens. HP delegators and IUC token holders also get upto 20% additional vote weight.
Here are some handy links for delegations: 100HP, 250HP, 500HP, 1000HP.
100% of the rewards from this comment goes to the curator for their manual curation efforts. Please encourage the curator @steemflow by upvoting this comment and support the community by voting the posts made by @indiaunited..
This post received an extra 20.00% vote for delegating HP / holding IUC tokens.
I'm doing the oppposite. I have some 3090's coming in a few days to build a dedicated AI server. I'm thinking about grabbing some A6000 though, as they perform slightly better than 2 3090's and have similar ram.
It really depends on what you want to do, if you want to be able to use a lot of models, and different tools, having your own hardware is key, especially during development. Deployment, you are almost always going to need to use cloud.
Hardware can give you access to near ChatGPT 4 quality models for relatively cheap. The next step is up is astronomically more expensive though. Dual nVidia 3090 (about $800-900 each on eBay) will allow you to use 70B parameter models, this is near top tier for open source models.
Macs allow you to get to the next step a lot cheaper, but with limits. The Mac Studio for example will allow for 192GB of unified ram which will allow you to run models with ~148G size for about $7000.
You probalby won't save on API fees (unless you are doing a ton) but you will have a lot more options.
Wow nice. We tried purchasing an RTX 4070. It was available on Amazon US but we couldn't get it in India.
https://www.amazon.com/ASUS-Zenbook-OLED-14-5-UX6404VI-DS96T/dp/B0BZ5P2DW3/ref=sr_1_3?keywords=asus%2Bzenbook%2Bpro%2B14%2Boled%2B14.5&sr=8-3&th=1
The fact that new things continue to evolve sometimes discourages me from buying things. When it is outdated, I’d feel like buying another one again and one can’t continue to do that
You may advice your brother to import the system
oh i will remember your advice.
This is really an eye opener that I really need to take careful note if as it is very important
It is quite astonishing to see the fact that things are really happening at a very fast pace and this is helping them