roguas

im moving my stuff as we speak, if those were just vms for services, no biggie but if you run business thats tied to repackaging compute - well, yeah there are better options

digitalocean has better pricing now, im actually migrating stuff right now, gcp also but only slightly

hn post = verify if you scale codeberg doesnt

i seriously hate my it dept attempts, they send you a link, you click, boom you have to enroll to a training

im sry, did i miss the part on how you can hack someone by simply sending them the link? is the web seriously that bad? honestly at least do full job and create some phishing website that goes along, otherwise wtf?

good "best laptop 2024" search results, absolutely exist, but they have to be heavily personalized towards user and i dont think we adjusted to such levels of in search (we have expectations of median/commonality there)

I didn't trust myself not to blow both my feet off with Clojure

But you are clojure engineer(by trade). Kinda weird, cause on the other end, I myself do backend in clojure but in frontend i much prefer js/svelte. (cljs tooling is making me cry, almost always)

No, open source means that sources are open, typically for inspection, modification etc. Also here it can be considered the case. Likely in order to claim "true open source", they would have to share dataset? But even this might not be enough for truely open source model? This dataset is nothing but another artifact. So how did they arrive at this dataset, now they have to share pipelines and infra...

.. the thing is, we have not dealt with llm much, it's hard to say what can be considered open source llm just yet, so we use that as metaphore for now

All the usecases we see. Take a look at perplexity optimising short internet research. If I get this mostly right its fine enough, saved my 30 minutes of mindless clicking and reading - even if some errors are there.

or internet consortium of financial advisor will tell you: it is a bit of your own fault, you should always diversify

we haven't had time to really adjust emotionally.

I think we adjusted, but we just apply broader context. No one would recommend going all-in on FTX with your life savings. Likely no one involved is really poor and this affected his life strongly. People lost money for their retirement retro vehicle, it's fairly easy to adjust to that.

This not saying it's ok, but its annoyance in effects mostly or should be(I think likely there are may be people, who put too much money, but then again internet financial advisory board all appears and says: HA! Told ya not to do it!).

Just stranded without dignity or purpose,

Perhaps dignity is seeking value that is scalable to entire humanity, hm? Isn't it dignified? If we cannot find value for fellow humans we can relax I guess?

We have outsourced art to computers, so people who don't understand art

Ok, I am seeing a lot of bad faith arguments. Just because somebody doesn't have time to learn to craft and cruft of some very tedious discipline, doesn't really mean they don't understand it.

You don't need to throw rocks all day to get an understanding of physics. But people had to do it in the past, now we calculate, simulate and then just do one throw.

Also I am speaking as part time designer/artist. New AI tools are great, I could achieve similar results, but difficult bets were made, when i started. Now I just roll in images, some of them finetuned on my work, some are mix, some are completely nuanced. Many artists after initial sobbing(I was enthusiastic from getgo) started to learn and explore this tools.

People bring amazing things to life, that just were impossible with their limited resources (take a look, anime done with very small crew https://www.youtube.com/watch?v=tWZOEFvczzA).

I much dislike the mindset of "they took our jobs".

I think the goal was last attempt at nation building in middleeast (that started in 70s, due to oil being way more important for US than now). In order to do nation building of course you have to shake the country and its structure, but you try to shake out the bad people, install stable and reasonable government and plant seeds for democracy, try and introduce some liberalism, education, infrastructure.

It failed, yes. But it succeeded in great many places, eastern europe, large part of south east asia, some african countries.

Russia is just a sorry ass loser, their country is in downward spiral and needs to exert power over its sphere of influence, thing is that countries in that sphere some time ago decided they dont really want to be aligned. Thus Russia decided to punish Ukraine otherwise the rest will also fall out of sphere (ironically I think they accelerated that process even more).

I was drooling each time he said Poland in that interview. Putin's got special kink for us. Its not well known fact but through weird tricks we ended up ruling Russia for a couple of years. Not a lot of folks on both side knew about it, but then Putin send money for a movie 1612, depicting their heroic retaking of motherland from polish oppressors.

I don't think Poland is very doubtful towards US(some is always welcome). We have strong ties and generally are on extremely good terms with US compared to other western countries.

As for weapons, well its a market situation, sometimes perhaps having non-us weapon systems is actually better.

Google and MS should offer support. While both of them operate app markets they have not went hardcore into ban all other markets, lock platforms.

To combine with previous "techdebt" mentioned. I dislike tech debt, as term because it causes false analogy. Prefer friction, which is how much harder/longer development becomes as compared to your original estimate.

While observing friction you get better idea and can consider actual cost of techdebt. There are many situation where you may care less for techdebt, you have a dev script that looks bad, but you likely just gonna replace it fully when time is right.

Yet friction is painful as it slows you. Debt analogy perhaps becomes more viable now, with high interest rates. This is what tech debt is, its not debt on 0% interest rates, its debt in high interest rate scenario, everything slows down.

How do you dismantle profitable business?

Hold on. I would avoid broad claims. Latest generation of automated phone systems is crap, yes. But we are very slowly starting to see something new here.

I can assure it would take me a week to fix a lot of problems aka memes coming from this. System prompt can be first place to start fixing, second small model or some another background call for just keeping conversation sane and within certain topic / rules (sort of like more independent conversation observer process to offload from original context), third you can finetune the model to have a lot of this baked and so on.

While this example is premature implementation, they are spearheading something and will learn from this experience and perhaps construct a better one.

yes, they dropped instruct model already and i think people are churning the variety as we speak

Euro folks here: Oh, this thing again.

Oh it's good they working on important problems with their ai. Its just openai was working on my/our problems (or providing tools to do so) and that's why people are more excited about them. Not because of cultural differences. If you are more into weather forecasting, yeah it sure may be reasonable to prefer google more.

Whats not possible? Do they ban people for exploring sota model that they offer?

Is it non-reproducible? Also results which reproducibility can be measured and appears stable is perfectly good science. I dislike when people throw statements like that.

Especially given this would be fine definition of engineering+science: "All you're doing is exploring the parameter space of particular model in a very crude way."

It may be a conflict of interest but it will be hidden by layers. OAI has deal with gpu producer and there is little conflict as they don't interact, but MS that owns Azure and resell of hardware creates that conflict. Maybe I have bad view, but I think many such conflicts that come from interdependent parties exist and especially in datacenter/cloud world.

Acknowledging a mistake so early seems like a sign of weakness to me. Hold the hot rod for at least a minute, see if the initial pain goes away. After that acknowledgement may begin to look like part of learning and get more acceptance, rather than: oopsie doodl, revert now!!!

I think AI, particularly text based, seems like a cleaner problem. Robots are derivative of AI, robotics, batteries, hardware, compute, societal shifts. It appears our tech tree needs stable AI first, then can tackle the rest of problems which are either physical or infrastructure.

Im swooning over gpt-based solutions most of my day (coding included). I consider myself professional, hope that other professionals are also swooning, cause it truly impacts at many levels.

a lot of disagreement,

pluggable llms is not something most people look in a product - its rare case of enterprises that are afraid of everything atm, also finetuning should be considered for costs not really range (i have not seen examples of finetuning that couldnt be achieved via prompting in llms, if someone has any, would be great to hear)

selfdriving cars are actually going for e2e systems now, there is still infra/architecture and perhaps even different models that comm with each other, but regardless system behaves as e2e solution that learns internal hierarchies of all the signals/tokens

large supersmart llms is where the game is currently at, they will get smaller, faster and more efficient we already are on this curve - if you're building llms i think you will fall behind

outer leafs of your org fueled by ai - while yes, this model is good for already established companies yet its also super important to think "ai first", ai much like any prev technology can be attached to legacy solutions making them better or can be shaped into a completely new unimaginable before solution... i wouldnt give vanilla advice that discourages the latter

Yes western companies spend a lot on marketing, cause without it you might confuse their products which are built to deliver positive experiences and value with similarly looking but not so positive counterparts.

Not to dunk on China particularly here, I do/did enjoy a lot of hq chinese products.