Great idea. I'm thinking if it could make sense to send the output to a cheap / local model to filter out only the bits that "matter" and pass that through - for the cost some extra time, but maybe it's worth it for saving tokens in the larger model.
It has nothing to do with the Industrial Revolution. Put yourself in their shoes. They are fresh entrants to a job market and these companies seem hell-bent on making them useless in it.
Which was a legitimate criticism and took decades of fighting to resolve through organized labor, and protests. There was no 5 day work week or laws against child labor or wage requirements or against nearly any exploitation of employees at its advent.
If AI is like an Industrial Revolution which is yet unproven, but is being marketed as such, then you would expect exactly this reaction.
Believe me, I get it, I get how scary it is, for fresh grads, for people with not a lot of work experience, to be bombarded by this stream of AI-centric talk. Even discounting what the hype men will have you believe, this technology WILL have a huge effect on employemtn in an information services-heavy economy. Many of the grads probably used AI themselves a ton to write their papers, which ironically may have increased their fear of the technology rather than making them see it as a multiplier of their own skills, as they saw first-hand how effective it is at coming up with ideas they struggled to articulate themselves.
Also implicit in my response was the assumption that college grads don’t generally boo their commencement speakers. Perhaps I am wrong, perhaps that’s pretty routine. And yes, I agree these are just platitudes, that have gotten very tired - but isn’t that how most commencement speakers are? Platitudes, feel-good stories, motivational cliches, all coming from people who lucked their way to the top of some pile and thereby qualified to be invited to graduation? To go back to my earlier assumption - surely the vast majority of commencements don’t lead to booing, in spite of how disconnected and distanced most of the speakers probably are?
I am genuinely trying to understand the nature of the polarized reaction to AI. Just from the perspective of its value to me, I feel buoyed by it, and not worried that it’ll depress my personal value, my financial future, etc. And yes, again, believe me, I get it - I get that I have a lot of cushions and privileges that allow me to feel that way. I am by no means imagining everyone else is in the same position.
But I’m really fascinated at how much more negative the reaction to this technology is in the US, vs. elsewhere [1]. Are Americans somehow more perceptive and alive to AI's danger, than, say, the Germans and Indians that seem to be more willing to find something to be “excited" about, with this technology? Others commenting here have said that the speaker's message, that this technology has just made their college debt much harder to work off, is naturally off-putting. Okay, I confess, I guess I just don’t see LLM AIs as representing such a significantly more dangerous threat than so many other factors that have roiled the American economy over the last 2-3 decades. Offshoring, rising costs, stagnant wages, loss of labor power, inequalities compounded by a severely tiered educational system, constant cuts in public investments in social services - these have been secular trends since the 70s. Have commencement speakers been booed all these years, every time they spout some boilerplate language like, “globalization is the new reality, let's get prepared for it,” or, “Finding your own brand is essential to participating in the 21st century economy”?
I think part of the answer lies in this quote from the NYT article [2]: “Earlier in the speech, Caulfield [the speaker] had lost some of the crowd by praising wealthy corporate leaders, including Jeff Bezos. It didn’t go over well. [A fine arts Bachelor’s graduate that day said,] “Using Walt Disney instead of Bezos would have felt generic to me, yet still demonstrated she understood she was speaking to a crowd of artists.”
There’s an aspect here I think not of opposition to the technology per se but what it says about which large-scale social investments seem to matter to those in power. I think it’s somewhat amusing, and telling, that this person thought to reference Disney of all people - someone that by no stretch of the imagination could be described as a model of corporate responsibility or of support for labor - because in this view, Bezos is so diabolical, even Walt Disney, by virtue of his creative talent, is a benign example of corporate success.
What “Bezos” represents is the idea that the financial and political elites would rather gamble the economy on multi-trillion dollar bubbles, that aggressively extract even more natural resources from our shared commons, or send the privileged few to gaze at this planet from the “edges of space,” than figure out how to keep, say, the cost of insulin low. That’s what got booed, I think, more than the what “AI” represents.
All the websites currently blocking Claude Code or other AI agents are fighting a losing battle. Computer-use is in the early stages, and the thing preventing mass-adoption seems to be the number of tokens it takes. Agents can fumble around trying 10 CLI commands that don't work before finding the right one and we barely notice. But other visual agents (browser use / computer use etc) end up eventually fumbling on to the right thing, but we don't have the patience to wait 20 mins. to click a button. As tokens get cheaper + faster, we probably get the models that can use a UI interface just as natively as a CLI.
Tokens cheaper? I don't think that seems to be the case ... VC funded tokens were there to build user base and token price will go up as they eventually switch from growth to profitability.
I wish I could place a lot of money on the opposite side of this bet.
I don't think many realize how could the cheap, alternative models are becoming. I prefer SOTA models for key work, but I can also spend 10X as many tokens on an open model hosted by a non-VC subsidized provider (who is selling at a profit) for tasks that can tolerate slightly less quality.
The situation is only getting better as models improve and data centers get built out.
Yes, but how cheap is it to run four at the same time? It’s tough to run one good model locally, but running four at the same time which I commonly do with Claude and Codex just doesn’t seem to be happening anytime soon.
I'm referring to hosted models such as via OpenRouter or from the model providers' own services.
I think everyone making claims that inference is getting more expensive are unaware that there are more LLM providers than Google, Anthropic, and OpenAI.
Fair - there are bets both ways though I wouldn't consider it to be a certainty. That revenue drive on this AI build out is going to be real and multifold.
It will take a few years until scheduled data center construction finishes, and together with software optimizations that may come up in the meantime, it may cause a significant decrease in token price.
And the lethal trifecta but I suppose that's all agents as of now anyhow. Every AI provider has major warnings about letting AI have access to PII in the browser.
If only General Relativity had such an ironclad defense of being as unfalsifiable as Foom Hypothesis is. We could’ve avoided all of the quantum physics nonsense.
it doesn't mean it's unfalsifiable - it's a prediction about the future so you can falsify it when there's a bound on when it is going to happen. it just means there's little to no warning. I think it's a significant risk to AI progress that it can reach some sort of improvement speed > speed of warning or any threats from AI improvement
If your revenue model is predicated on Star Trek-style communism, it's maybe not a very realistic model. I really don't think if producing things is essentially free that advertising will be a very big thing since it would be pointless.
I was working on something almost similar, mostly to create a browser agent, and LLMs are really good at writing playwright scripts compared to vanilla JS.
I didn't realize the setup script had to be done in the UI over in the environment tab. I assumed it would be reading something like setup.sh from the codebase.
We're looking to build a team of experienced software developers to help us bring new products to market in cross-border ecommerce space. Our stack is based mainly on a Ruby backend and we have a number of different technologies we use at a smaller scale. We deal with directly with the logistics, e-commerce, search, and web crawling. There's tons of interesting opportunities in the space and we love to move fast. Also happy to sponsor visa / help with relo.
Email me if you're interested, or want to chat further: rahul [at] [companyname].ae or apply at link above.
Open Positions:
- Senior software engineer (Ruby/JS/React experience is a plus)
- UI / UX designer
- Product Manager / Product Lead (experience in ecommerce / logistics is a plus)
- Frontend dev (React Native experience a plus)
