Ask a career advisor: how mid-career people can help with AI and how to survive a job hunt

This week, we’re answering our newsletter subscribers’ career questions.

Question one: I’ve been working over 20 years as a software engineer, and participated in a few publications as a research engineer. I’m open to getting deeper into research, in fact I’m starting a PhD in computer science. I’m close to 50, but most of the career advice I see from 80,000 Hours seems focused on younger folk.

I’m wondering about reorienting towards global issues; does it make sense to continue with my current PhD? Drop it for one directly related to (e.g.) AI? Drop the PhD and just apply as a research engineer to appropriate companies with such engineers? Or just leave it to the next generation?

Sudhanshu: I would pretty much never say “leave it to others” — there might not be many others, and experienced folks like you often have a lot to offer. We are always hearing from impactful organisations all the time who want experienced, mid-career people to join them in key roles.

With 20 years as a software engineer, I’d be excited if you could apply that experience directly to a pressing world problem, unless you’ve gotten some evidence to suggest that you need some new skills. I think it makes sense to try your hand at just working directly on the problem you think is most important before moving on to other options.

If you do think working on reducing AI risk is the most important problem, as I do, I think you have lots of options. You might find opportunities to do this work in your PhD, but it’s probably cheaper to test out your fit for high-impact software engineering work. This could look like:

• Doing empirical AI safety research as an engineer — for example, on open source projects like EleutherAI, or this software engineering role at the UK AI Safety Institute
• Working in security for powerful AI models — like this Director of Security Engineering role at Google DeepMind
• Working for high impact nonprofits developing tools and technology — like this member of technical staff role at Sage

For many more ideas, read our advice on software engineering and AI safety technical work, find relevant jobs on our job board, and check out these projects, guides, and pieces of advice that you could work through.

Some of these options are potentially easier than a multi-year program, like a PhD. But you don’t have to close the door on that — spend some time with the above resources, and you can stick with the PhD if that doesn’t fruit.

Question two: It’s disheartening to have so many job postings that don’t post salary, yet require so much time to apply. It feels like I can spend loads of time filling out an application, including case studies, answering questions, etc. — and then get a low-ball offer. I never would have applied in the first place, and this feels like a problem everywhere. How do you get past this? It just feels like why try unless you’re out of a job.

It’s so “easy” to apply with tools like Indeed’s auto apply, or by using AI, but then you get no results. You spend a long time trying to make a stellar resume, cover letter, etc. and then get no response. Adding my “voice” to cover letters never gets a reply back. It’s like people need the general, plainspeak to have your resume looked at.

I guess my question is: how do you get past the bleak, bleak outlook of even trying to apply in the first place. Everything seems so stacked against you that it’s hard to envision success. Networking seems like the only answer, but with a gut-check of ‘ew’ and a recognition I don’t have a network. I don’t like social media, and need to get a LinkedIn going, and that’s about all I can think of. The whole topic is depressing and disheartening.

Sudhanshu: Job hunting can certainly be an unpleasant experience, and unfortunately, there are no simple solutions to the cluster of questions you raise.

Regarding salary offers, it’s almost always true that the first offer on the table is the lowest a company is willing to pay without being truly exploitative. So you should probably expect it to be a low-ball. While the strategy of companies being cagey about their salaries might have made sense for them in the past, with so much publicly available information (e.g. Glassdoor), it seems silly today. I’d encourage you to do your due diligence by sourcing reasonable estimates from the internet on salaries for the roles you’re applying for, in that company and also possibly its competitors.

When you apply and don’t get a response, it doesn’t hurt to write back to the company nudging them for a reply. I’ve heard more than a few stories where this demonstration of agency was rewarded with feedback and sometimes forward movement of someone’s application. Of course, it might be prohibitive to do this for a hundred companies, so pick just a few for such targeted outreach.

I’m glad you noticed your own gut reaction to networking and the signal it carries. However, in this case — as is the case with so many other things, like green vegetables and exercise — it might only seem unpleasant, but actually be really good for you and become more enjoyable over time. Networking doesn’t have to be about gross status-seeking — it’s actually a productive way for people to match the skills they bring to valuable opportunities that others have to offer. Despite your worries, I think you should try it out.

But trying a new thing can be daunting, especially if you’re not immediately excellent at it. I find that it helps to suspend judgment for the first several hours of attempting a new skill. Most people see rapid improvement within the first 20 hours of practice and their inner experiences are qualitatively different than when they began.

All of which is to say — don’t take any paths to success off the table prematurely! If your own network seems thin, create new ones: attend events in your field like meetups, conferences, talks, or hackathons regularly so people get to know you and learn more about what you’re capable of. Send cold emails like a boss.

Also, think about how you can make your skills legible in public. If you’re a programmer, a LinkedIn page is much weaker than a GitHub that actually shows you can code. If you do research, publishing something on a public platform like the Effective Altruism Forum and engaging with the comments can demonstrate greater clarity of thought than a few disjointed notes on a hard-to-find personal blog.

For more tips, check out part 11 of our career guide: All the best advice we could find on how to get a job.