> any sort of LLM writing is a tacit acknowledgement that people don’t really care what they are writing Couldn’t be further from my experience. I work with a distributed team that has four different native languages, with substantial engineering staff for each. LLMs have made our cross-geo communications so much better. I can get a summary of a standup that happened overnight (to me) and in a foreign language (to me…
How accurate are the summaries?
Based on my experience, it's very risky. A lot of times, it connects similar but very different things that should be seperate and changes the meaning of the message/discussion
Here is what I want from Apple Intelligence: Given a user story “I chipped my tooth” I say “hey siri, I just chipped my tooth, can you help me?” I would want it to: - help me schedule an appointment with my dentist (it could create a reminder, ideally know who my dentist is, even ideally queue up or place a call) - put that appointment on my home and work calendar - remind me to tell my manager I need the time away -…
Ohh and I would like to be able to tell it not to keep featuring photos of my ex wife years after the divorce.
Google photos somehow gets this right. Doesn't feature any pics from my previous relationship.
I’ll tell you why this happens. You might use ChatGPT for a bit and your initial impressions will be great. It does what I ask of it! You might be aware that it makes mistakes sometimes, but when you use it, you don’t notice it because you’re using it interactively. Now if LLMs are just effective as your experience says, they are indeed extremely useful and you absolutely should see if they can help you. It’s only wh…
This definitely resonates but I'm left wondering why there hasn't been a collective "sobering up" on this front. Not on a personal/team/company level, but just in terms of the general push to cram AI into everything. For how much longer will new s assault us in software where it ostensibly won't be that useful? It seems that the effort required to make an LLM work robustly within a single context (spreadsheet, worddo…
Very few managers get quick promotions for NOT rolling out a high-visibility AI enhancement. LLMs can theoretically fit into an amazing diversity of products. Even if just 10% of managers say yes and the other 90% say no, thats still a lot of shoehorning every year in an attempt to book a “win” for a promotion.
An abstract syntax tree is not semantics. And language models don’t do this kind of explicit syntax parsing and representation at any rate.
I'll admit I haven't seen the training data but some basic googling shows a subset of the labeling is syntax annotations. I am not claiming LLMs parse code in the way you are suggesting, but they certainly have token level awareness of syntax and probable relations which are the roots of any programming language.
Last time I researched AST parsing was quite rudimentary for LMs. The problem was preserving their structural properties, whereby a flattening approach via node traversal tended to remove. But you needed to do to put it into a format that language models could parse.
This definitely resonates but I'm left wondering why there hasn't been a collective "sobering up" on this front. Not on a personal/team/company level, but just in terms of the general push to cram AI into everything. For how much longer will new s assault us in software where it ostensibly won't be that useful? It seems that the effort required to make an LLM work robustly within a single context (spreadsheet, worddo…
I can tell you that there has been a lot of sobering up — but that the news isn’t made by those people…
Totally. And everytime someone sobers up, there is a cabal of people saying "we've sunk however many $$$ into this, it's the core feature of the xx roll-out...drink up, the hype party continues, like it or not...." So now you see phenoms like the one-time 'premier tier subscriber only feature of co-pilot on Github now pushed to everyone, prompts to use the generative A.I in iStock on every page, compulsory "use Co-pilot to write your draft' prompts on every new doc in MS Word - because I don't think companies are able to grok the widespread disinterest in much of it. I'm still waiting for one that will be non-networked and sit on my desktop to do my tax returns and haggle with phone company bots.
> Ask it questions about accounting, other domains it knows Be very careful here if you're using it for anything important! LLMs are quite good at answering questions about accounting in ways which are superficially convincing-looking, yet also complete nonsense. "But the magic robot told me it was okay" will not fly in a tax audit, say.
Exactly my immediate reaction. Accounting has to follow very strict rules and needs some application of judgement. It might answer questions in a useful way, but you have to make sure you understand the answers and that they match accounting standards or tax rules (and one danger, at least in some places, is that they are different and you might apply the wrong one).
Unfortunately everything I've asked any of the main LLMs about where I actually legit knew the precise answer, they were wrong or half-right but excluding important context.
I was reflecting on old technologies that were supposedly revolutionary at the time, like palm pilots, and what value they really provided. Like, I can imagine a someone at that time using a palm pilot and claiming that it's making them more "productive." Looking back though, such technologies are nothing compared to what we have. So there's a perceived gap between our current state of technology (LLMs etc.) and the…
The productivity improvement is negative because of how much of everything is now a "product" instead of just a tool. Whenever you open something, especially after not using it for a while, you're met with an obstacle course of in-your-face modals and popovers about new features that you don't give a shit about. Software used to let you get your job done more efficiently. Installing a new app and learning to use it f…
This! Rather than focusing on a robust, bug-free "it just works" product - the original Steve Jobs vision embodied in late 90s/early 2000s Apple advertising, the offerings are desktop computers with hundreds of "cool" things you can do, like we're all teenagers. And once you click away all the pop-ups and disable the unwanted features, you're left with software that has a bug in it that's precisely the same as three versions ago. As for Apple Intelligence, I think I am really weird but when I am trying to find 'the answer' to something, I find the research process super informative and always get fresh ideas or new, peripheral input as I go.