Yes. I've always suspected inference is profitable.
But I'm also interested is retraining needed? how relevant will a model be in N years if it's just "static", and if you do need to retrain and fine-tune over time, is it still profitable? how often is that needed?
For most cases it's probably ok, often times answers are general or basically a web search anyways, but if my Agent started blasting outdated syntax it would suddenly be way worse for coding for example.
They need to be retrained often is my experience for general usage.
I have SO many cases, for example when shopping, when they make assumptions based on their cutoff date, and they need to redo everything based on lots of searches.
Sure, for many use cases old models will be useful but for general usage a model that's 5 years old will probably be garbage for most people, unless they prompt it extensively, which the general public won't do.
Was this a common thing? I don't know which bubble is more common, but personally I can't recall ever working with someone who couldn't explain what they did (just copy paste?), except for maybe configurations.
In my experience, people most definitely understand less of the technical implementation now vs before. Heck, not only technical, I've seen presentations from business where the presenter clearly vibed their deck and don't understand the content.
It was a huge thing, and I lived through all of it, yet never ran into a problem where copy-and-pasting would have been the lowest resistance solution. I never even ran into a problem where asking the question and waiting for a response would have made more sense than asking internally or working through it myself. The most difficult problems are relatively conceptual, which SO didn't like anyway.
I never really got it. I think it depended bigly on which domain and stack you were in.
My bubble is opposite - majority did not understand simple things. I ve seen people randomly trying SO advice on “fixing” git repo, mutating objects passed as arguments (because it works!), not understanding what is env variable.
Your example just proves the point: AI is not a problem - it’s a leverage. Give it to someone who don’t care - they will do as much damage as they can.
But imho the issue is not ignorant people. It’s lack of leadership. Somehow those people got selected for the job and still keep it.
I think it's both. I want to agree with you but it's also just human nature, I notice myself slacking even when I care, because it's easy. The human brain is really good at finding shortcuts.
It's like saying it's the kids fault for being addicted to social medias and assigning no responsibility to the providers.
To be clear, I'm not advocating anti-AI laws or arguing with you, just blurting out my thoughts. Ultimately I simply feel this will not be a good thing for tech and human understanding in the short / medium term, but this is just based on vibes, like most arguments on AI are.
I think it's fairly likely medium to large corporations don't pay anywhere close to the listed API pricings as they get deals through existing partnerships with the big cloud providers.
The free models for ChatGPT, especially without login, do very little reasoning. You should at least log in to set any level of reasoning above Instant, which uses virtually none.
Is this something that happens by accident in the sense that the decision makers didn't know how it would sound or is there actually someone in a room at some point going "fuck it".
But hasn't bio attacks already been possible before AI, what has changed? Maybe the barrier to entry lowered, but one could argue the availability of info has not been the bottleneck for as long as the internet has existed.
The threat is ofc real, but AI won't magically "do the thing" still, it's not code that's the bottleneck AFAIK? Correct me where I'm wrong.
Synthesizing complex information from a range of technical sources enough to build bioweapons is cognitively impenetrable for the vast majority of people. The ability of AI to synthesize this information into a step-by-step recipe to follow is a step change in accessibility.
There's an opportunity cost to terrorism just like with any time sink. The effort to build up knowledge enough to produce some weaponized pathogen will be compared to just doing traditional terrorism. For some low capability terror cell, its easy to see how the cost/benefit analysis has been in favor of traditional terrorism up to now. The kinds of terror acts that take years of sustained effort to execute are rare. But as the barriers to entry to bioterrorism fall away and become widely accessible we may see the cost/benefit shift.
> Synthesizing complex information from a range of technical sources enough to build bioweapons is cognitively impenetrable for the vast majority of people. The ability of AI to synthesize this information into a step-by-step recipe to follow is a step change in accessibility.
I agree with this part.
> But as the barriers to entry to bioterrorism fall away and become widely accessible we may see the cost/benefit shift.
But I argue that the barrier to entry is not about information, it's about needing to build, execute and maintain a lab and production facility? It's not the theoretical part that's the bottleneck.
Or maybe I'm overestimating the difficulty of implementing such a thing, this is not in my area of expertise.
>Or maybe I'm overestimating the difficulty of implementing such a thing, this is not in my area of expertise.
Not mine either. But there is a vocal contingent of experts in biotech that have been sounding the alarm for the last couple of decades at the potential for rogue groups to develop weaponized pathogens for terrorism. And this was years before LLMs. Apparently constructing a sufficiently capable lab isn't really a bottleneck due to how freely available equipment is by mail order and how unsophisticated the processes are in practice.
>Consider our publication earlier this year, in which we reported that 36 of 38 companies shipped pieces of the 1918 influenza genome. That was collectively enough to generate the virus using reverse genetics several times over – and they were ordered by a student using a pseudonym falsely claiming to work at an office address. No one asked whether we had permission to work with 1918 influenza (a select agent), or why we needed genes from it, even though the official U.S. guidance at the time instructed them to do so. In fairness to the companies, that guidance was voluntary, and the fragments we ordered were legal to produce and sell domestically in every nation of the world (export controls are complicated). That is still true today, which is why we have a problem.
In stuff like wikis and ticket systems what I see is people now just making their own little tools around the ones we already use, this arguably just sements their position even deeper.
Also I dont get the inhouse argument, you dont really save anything. So instead of paying X to SaaS, you now propose to pay X to anthropic instead AND maintain said SaaS... what? how is that better, that just shifted the cost to another company and you got the burden of maintenance and responsibility.
What I mean is that a lot of the popular languages have an engine written in C. Fairly common pattern. I get it that we want to abstract but sometimes it feels we just want to avoid dealing with a language that is one layer below.
I agree with what you said, and _sometimes_ it's true, but I still wouldn't call it a "preference". One could argue there isn't a "need" to deliver faster and come up with solutions to problems, but is it a "preference"?
Maybe I'm completely off base here, arguing about semantics.
I remember fighting with business about this in a startup way before LLMs, more code, velocity or motion doesn't mean more value.
If I put on my optimist hat for a moment, I hope LLM's would finally make this obvious for people, more stuff !== more value. 250k lines of code a week isn't a flex IMO, it doesn't really mean anything out of context.
But I'm also interested is retraining needed? how relevant will a model be in N years if it's just "static", and if you do need to retrain and fine-tune over time, is it still profitable? how often is that needed?
For most cases it's probably ok, often times answers are general or basically a web search anyways, but if my Agent started blasting outdated syntax it would suddenly be way worse for coding for example.
reply