techpression

I actually have very detailed specifications around the CSS, the problem is that LLM's are trained mostly on codebases with high amounts of complexity so they like to create layer of abstractions that only make your codebase much harder to reason about, machine and man alike. This is very visible in CSS, where almost every single thing automatically becomes a component wrapping the CSS (be it React or LiveView in my case), which is very often meaningless because the CSS itself is the component.

I love that we have this on one hand and me cleaning up catastrophic CSS made by Sol on the other. Then again, maybe CSS is the ultimate benchmark.

Thank you. I was starting to think the history revision was almost true, but your recollection is very much in sync with my own. Everything was machine learning, nobody talked about AI unless it was for research, now marketing has changed that, unfortunately.

That left side border is a plague…

The copy is usually a bit of a giveaway too, ”explore your future path to greatness and experience the defining divider between those who can and those who can’t” or as most humans would write the header ”plans”.

It’s not just about reading the screen, it’s about using it. It’s quite the downgrade going through a slow API hoping to be able to navigate compared to the 10x speed (or more) of screen readers. Then you have the problem of LLMs being stateless and sensitive information etc etc.

I fully expected this to be complements for the output you get from those tools for some reason, that AI generated look. No idea why anyone would want that though.

Thanks for the link, it seems a lot more capable and interesting, to a much better price.

Ah yes, I very much agree on that framing.

The calculator comparisons are truly meaningless imo, a calculator does nothing if you don’t know how to use it and what to input, an LLM circumvents all that, but a lot of people seem to think it’s the same.

It's also very very far from anything general intelligence. This is actually where I think we see the best ROI, we have a specific task or problem and we target it. I remember reading about a model trained on railroad tracks. By putting cameras on trains to take images while running like normal, then using the model to detect cracks before they cause issues.

Why would we be able to design a safety plan that would control such a powerful force? It would be like putting umbrellas on an asteroid hoping to slow its fall. That sounds like delusions of grandeur.

Has anything really important been solved by AI yet, or where is this radical (imo) belief around AGI coming from? Genuinely curious, I know there are some math problems solved and ML has been used for far longer than AI to improve things, but where is clean (efficient) energy, the cure for cancer (or any of the horrible neurological disorders, take your pick), new hardware designs, quantum computing solutions, etc etc, you get the gist. Where are the things that will actually send humanity into the next era of civilization, I don't care about more React apps (but I do enjoy my coding companions for other things).

Heck, a proof for P=NP or P!=NP or solve the The Riemann Hypothesis. Just give me something truly exciting and I will believe AGI is around the corner, until then I will see it as cool technology, that while beneficial to me, also helped cause the biggest amount of disinformation we've every seen.

So company A pays for your education to go work for company B (their competitor) and all they get is a lousy interest rate with no security? If they don’t get ten years (or whatever amount) of your output, why waste the time and energy when someone else will do it for them?

And now we’re desperately trying to ”upgrade” penicillin (and friends) because it doesn’t work any more in many cases. Do you think we can repeat the process or do we need something completely different?

This is why biological comparisons are weak, we talk about a few agents verifying and checking LLMs, meanwhile the world consists of almost an infinite number of the same, just operating on different time scales. I agree that with we don’t know the timescale, and we definitely don’t know if long term it will continue to work ”adding more of the same”. Throwing more penicillin at the problem sure as hell didn’t, but it looked great initially. And I’m obviously not arguing the human benefits of penicillin, just that what we thought would work forever quickly didn’t.

Same thought I had, I realized I was zero percent surprised reading the claims made, it feels like a perfect representation of the personality Sam Altman shows the world.

Don't forget price fixing [1] which we are seeing clear indicators of happening right now too.

[1] https://en.wikipedia.org/wiki/DRAM_price_fixing_scandal

I would see less fake news, fake profiles and fake content in general. I would be happier, even if I would miss some of it. It’s kind of sad how much we accept the idea of ”trust absolutely nothing” nowadays, even movie trailers for fictional movies are made to drive clicks for ads… obviously there has always been a large trust issue online, but with gen ai we entered a new era of it.

The illusion of safety.

”Look the type says the data has this shape”, but it doesn’t. This has led to so many cases of ”no data validation” in my experience, how people solve that is usually then with zod, and now you have added a massive runtime dependency instead of local validation.

I've seen TS hurt more than help in many cases, because people think types define the world, and then they get malformed JSON (or just a new structure) and their world crumbles.

Yes, it's a skill issue, but oh dear the amount of developers who have that issue.

They've been doing this with unpopular votes since the inception of the EU, nothing new and people definitely haven't woken up, unfortunately.

There's a difference between standing in the rain and being hit by a tsunami, even if it's both just water.

Which is why we’re putting our entire digital identification infrastructure in the hands of Google and Apple. EU technological sovereignty is a kafkaesque affair, and that’s putting it mildly.

I think developers = studio in this case, management included. It’s the same issue as Hollywood really, making content for a loud minority will not generate revenue. Compare how Marvel Rival performs to any Marvel Hollywood content after Endgame.

Not at all the take I got, it was a specific, very corporate, section about ”say nothing bad about LLMs” that was later removed due to backlash that caused the author to lose trust. They even go into specifics regarding the LLM policy and that it’s in many parts sound.

When I read the mentioned section I was very baffled, it’s like if Jensen of NVIDIA had made an appearance and wanted to sell GPUs.

I love that quote, especially considering the insane amount of bugs that are produced. It’s as easy to debunk as someone claiming ”I can jump to the moon”.

They can’t keep their current models working on subscriptions[1], so we’ll see if this is marketing or not in the future. It’s smart to tease it no matter what, ”insert classic first hit is free drug reference”.

[1] https://status.claude.com

This is quite terrifying to me, because I have a feeling I will soon come to the same conclusion. I’m starting to see some really glaring omissions in code I’m responsible for (using Opus) that at first (and second) look seemed fine, but really isn’t.

I’ve worked at many companies where this idea of velocity was claimed to matter, and it never did. The only thing it mattered for was to make it look like middle managers were worth anything, but the success was always in the foundational idea/concept.

The graphics are amazing, feels Ghibli-esque, which is crazy to say in the context of a rich text editor.

That’s quite insane, especially considering how Google is pushing Gemini into every single product.