Page 43 of The Love Hack

‘I hoped you’d say that. Thanks, Lucy. And – you know – onwards and upwards, right?’

‘Onwards and upwards.’ It felt more like a dismissal than an encouragement. ‘But, Greg?—’

He’d already looked away from me, down at his screen, but now he looked back, his eyebrows raised enquiringly.

‘What if it doesn’t work?’ I asked tremulously.

‘Then we’ll need a bit of a rethink.’ As if to lessen the harshness of the reality, he produced an encouraging smile. ‘But we’ll make it work. Chin up, Lucy. We’ve got this.’

There was nothing more to say. What on earth had I let myself in for? Still clutching my notebook, which now bore a soggy handprint, I left Greg’s office at a purposeful walk – even though what I felt like doing was running away screaming for help like the building was on fire.

SIXTEEN

At home that night, Astro and I had a working dinner. At least, he tried to eat the scrambled eggs off my plate and I tried to work.

My meeting with Greg had jolted me. I knew he hadn’t meant it to – I knew he’d meant to soften the blow of Ask Adam’s disappointing hit rate, come up with a solution, reassure me that if I produced just a little more content, it would all come good.

But I couldn’t share his optimism – if it had even been genuine, which I doubted. The new-found confidence I’d been feeling at my ability to be Adam, to answer men's problems in a sensitive yet forthright fashion, had been dashed. The numbers weren’t stacking up. I wasn’t doing a good enough job.

I wasn’t good enough.

My heart sank as I flicked through Adam’s inbox. As I’d said to Greg, there was no shortage of material there to work with. Young men, older men, single men, married men. Sad men, angry men, lonely men. Men who seemed to hate women and men who seemed to think women hated them.

There were so many of them. They all needed my help, and right now it felt like I couldn’t even help myself. Back when I’d embarked on this crazy scheme, I’d had Amelie as back-up, but now I couldn’t even turn to her.

‘Come on, Lucy,’ I said aloud, repeating Greg’s words to myself, ‘You’ve got this.’

I opened the document where I’d saved the anonymous letter Adam had received earlier. At the time, it had seemed straightforward to answer. But now, my self-belief at rock bottom, I had no idea how to respond to whoever he was.

End things with the girlfriend, don’t end things? Hold out for someone you feel really passionate about, settle for someone you think is good enough. What was the right answer? Two people’s happiness could depend on what I decided to write – and never mind that, there were scores of other men whose problems hung in the same balance.

‘I don’t know if I can do this, Astro,’ I said.

Astro blinked slowly at me, his eyes like amber traffic lights.

And then I remembered Nush’s hissed aside during Zack’s wedding speech, telling me that he’d got GenBot 2.0 to write it for him, and how that had made me feel – not shocked, exactly, but kind of let down and disappointed that he couldn’t have been bothered to do it himself, but had used a generative chat algorithm to do it for him. It was his wedding reception, after all. He literally had one job.

I literally had one job, too. Right now, though, I didn’t feel like I could do it.

You can’t get artificial intelligence to write Adam’s column! Part of my brain recoiled from the idea in disgust. It’s cheating.

It’s not cheating. It’s just getting help, same as Amelie helped me in the beginning.

It’s probably copyright infringement. You’ll get sacked.

I’ll get sacked anyway, if I don’t come up with a way to answer these damn problems.

You’ll get sacked in disgrace.

I won’t use it to write the actual answers. I’ll just ask it to give me a steer. It’s no different from getting a steer from Chiraag before I answered Mark’s letter.

It’s lots different.

But I opened GenBot 2.0 anyway. I knew what it was, of course – I even understood the technology behind it, because I’d researched it back in the happier days when I’d been technology editor of Fab!. I’d written an article about five ways it could help Fab!’s readers in their daily lives – from composing tricky letters to their line managers to highlighting the top summer fashion trends.

I knew it was a generative language-based chatbot, which had been trained on a vast input of text which it drew upon to produce language of its own. I had a vague idea that it could be trained still further by its users asking it follow-on questions that would refine the output it gave not only to them, but to other users asking it similar questions. I’d delved into the theory behind how it worked – I vaguely recalled a load of stuff about machine learning, reward models and something called Proximal Policy Optimization, which I’d attempted to get my head around at the time but certainly couldn’t understand now.

But, although I’d tinkered about with it a bit, just to make sure it could do what I was saying it could, I’d never actually used the technology in my own work. I hadn’t needed to – until today.

---