What I find interesting is that Codex did everything except the one page even a research office couldn't have written for you. And I don't think that's because it doesn't know you well enough yet. A vision statement is supposed to be arguable, you're putting your name on a bet that could be wrong. A model trained to never rub anyone the wrong way just can't do that. Put two of them together and it's worse, because they talk each other into the safest version :)
I'm a product manager in AI, and this is the pattern I see all the time. The demo looks magic because most of the work has been done before. What it can't do is supply the nerve. Good that the nerve is still yours. Good luck with the ERC ❤️
I mostly agree except 1- I think I could get AI to write the vision statement by feeding it enough info, but mine is faster and better. 2- The LLMs are far from polite to me and to each other! The preliminary “interesting!” comment is a prelude to some pretty savage stuff. I have set both to no-bootlicking mode, btw.
"The heady feeling of revenge against bureaucratic gatekeeping felt like an irresistible surge through enemy lines."
I've always felt like AI taking care of the procedural and repetitive is inherently anti bureaucracy and giving more power to the interesting or even 'soulful' parts, but I find it hard to explain to the people around me who (legitimate job loss/instability fears aside) often take the stance that it's undignified to use AI.
Yes maybe ——- those that had servants — how many employ servants nowadays? Or office assistants?
I get yr point that there’s inherent resistance to new ways — it seems after the proving period there is either acceptance or not.
I’m still on the fence because it takes some time to understand the actual repercussions of epoch changing tech. And many times those changes are irreversible.
Same realization, different slice of the world- I've worked with labor unions for about a decade of my legal career. We were always the scrappy underdogs against huge corporations with sophisticated legal departments completely devoted to undermining collective bargaining. AI levels the playing field and allows the merits of the dispute to lead the fight, as opposed to the sheer manpower. I get a lot of disdain for it, but I am excited for the opportunities AI can bring, if we can just get out of our own damn way.
And btw lawyers, esp. in case-law countries like the UK, should be quaking in their shoes. Here in the UK a supposedly posh firm will ask for £10k on account just to write a stiff letter. In a recent dispute, I passed on to my attorney some really good, punchy, tightly argued, chapter-and-verse stuff from Claude and saved myself a couple of grand.
As a US lawyer, I agree. The AI legal platforms we have now should replace at least half of us in a matter of years. The trick is to be in the right half.
What I find interesting is that Codex did everything except the one page even a research office couldn't have written for you. And I don't think that's because it doesn't know you well enough yet. A vision statement is supposed to be arguable, you're putting your name on a bet that could be wrong. A model trained to never rub anyone the wrong way just can't do that. Put two of them together and it's worse, because they talk each other into the safest version :)
I'm a product manager in AI, and this is the pattern I see all the time. The demo looks magic because most of the work has been done before. What it can't do is supply the nerve. Good that the nerve is still yours. Good luck with the ERC ❤️
I mostly agree except 1- I think I could get AI to write the vision statement by feeding it enough info, but mine is faster and better. 2- The LLMs are far from polite to me and to each other! The preliminary “interesting!” comment is a prelude to some pretty savage stuff. I have set both to no-bootlicking mode, btw.
"The heady feeling of revenge against bureaucratic gatekeeping felt like an irresistible surge through enemy lines."
I've always felt like AI taking care of the procedural and repetitive is inherently anti bureaucracy and giving more power to the interesting or even 'soulful' parts, but I find it hard to explain to the people around me who (legitimate job loss/instability fears aside) often take the stance that it's undignified to use AI.
People thought answering the phone was undignified.
Yes maybe ——- those that had servants — how many employ servants nowadays? Or office assistants?
I get yr point that there’s inherent resistance to new ways — it seems after the proving period there is either acceptance or not.
I’m still on the fence because it takes some time to understand the actual repercussions of epoch changing tech. And many times those changes are irreversible.
On the fence on which point?
Unleashing the Kraken of AI ;-)
Attention derrière toi!
Yep been rocking Codex since its release early this year (for the traditional use case of software creation).
It's an extremely impressive piece of productivity software that skyrocketed my output. GPT-5.5 is bewilderingly powerful.
Same realization, different slice of the world- I've worked with labor unions for about a decade of my legal career. We were always the scrappy underdogs against huge corporations with sophisticated legal departments completely devoted to undermining collective bargaining. AI levels the playing field and allows the merits of the dispute to lead the fight, as opposed to the sheer manpower. I get a lot of disdain for it, but I am excited for the opportunities AI can bring, if we can just get out of our own damn way.
And btw lawyers, esp. in case-law countries like the UK, should be quaking in their shoes. Here in the UK a supposedly posh firm will ask for £10k on account just to write a stiff letter. In a recent dispute, I passed on to my attorney some really good, punchy, tightly argued, chapter-and-verse stuff from Claude and saved myself a couple of grand.
As a US lawyer, I agree. The AI legal platforms we have now should replace at least half of us in a matter of years. The trick is to be in the right half.
Why stop your analysis here, though? Why not explore what your last sentence really means?
Do tell..
I'm saving it for my (eventual) page. :)
That is very cheerful news, another brilliant instance of disintermediation. The disdain is extraordinary!