Show HN: Simple and fast resume document generation with AI | Hacker News Reader

Show HN: Simple and fast resume document generation with AI

cvgist.com

52 pointsprodtorok44 comments

Hey HN, my brother put a lot of effort into releasing his first app ever. Feel free to show some love on the product launch as well: https://www.producthunt.com/posts/cvgist?utm_source=badge-fe...

Feedback welcome of course!

Consider adding a FAQ on the website.

-----

I have an 8-page long résumé with good details about all projects I have done. Would it be able to shrink it to a two-page résumé?

Is there a way to enter additional prompts to customize what's desired in the output?

Are multiple runs allowed if the output is not satisfactory in the first shot?

Thanks.

Thank you, will do.

What're the answers to my own questions though? :-)

Is it ok to email you? We will set you up with an account, and you can try the resume shrinking for free. Sound good? We’d like to know how it works as well!

Sure. My email ia in my profile. LinkedIn too. Thanks

Sweet, I sent a LinkedIn invite

We wrote a blog based on the discussions here.

https://blog.cvgist.com/2024/05/23/ai-and-resumes-cutting-th...

It would be great if you had the option to get the data from LinkedIn. Typing everything out is a chore when it's already typed.out somewhere else.

I think I would be more likely to use this if it were a subscription model despite usually hating subscriptions.

Looking for a job is a time bound process and I want to know that I have full access to the tools needed for the entire time it takes to find a job. During my most recent job search the 15 rewrites may have lasted me less than a week. I was using ChatGPT to tweak every resume I submitted.

Thanks for the feedback. Yea, the rewrites are needed. Users don't always know what they want right off the bat, and it hits them later on.

Our original goal was to help people get their resume going and into a nice-looking doc asap.

College students we interviewed were adamantly against subscriptions, so we pivoted. We realize it's a more effort for seasoned professionals.

We're going to increase the cheapest offering to at least 2. GPT-4 has some cost limits, but we're actively figuring out how to squeeze more value for users. Our dream is to "speak to your document" and we're evaluating if we can and should try to get there. We think we have innovations in the word processing space... obviously it's not the safest area to play in with MSFT looking to innovate their own products with Agents.

I hope it’s okay that I am contacting you regarding a bug here. I purchased the standard package that offers 7 resume creations but after completing the purchase it says I only have 5 available to generate.

Sorry for the trouble! Fixed now. You should see 10.

thank you!

Honestly, don’t base your main pricing on college students. Maybe have a deal if they sign up with their college email address. For the rest, price for the customers you want.

It’s a great idea. I could also see pricing for professionals that are creating many iterations for many industries and roles. Something in a larger bulk, but favorably priced

For me, it is the other way around. I prefer a pay by use model.

I like the price point - small enough for an individual to commit.

Maybe you can try A/B testing to see if $2.99 does much better than $3.00 :)

I mean this question with the best intentions:

Can you share why you think this is a good idea and what do you think is the long term implications of AI-ifying resumes?

The optimist in me hopes that this sort of thing kills resumes as a major communication channel for good. High-volume, low-signal, wildly-inflated-claims-as-the-norm structures have been devastating to good hiring, and I have some hope that as gen AI gets traction in job hunting that it'll take resumes from "low signal" to "no signal" enough to force a change.

The pessimist in me says that that just kills inbound entirely and we go back to an almost purely connections/network-driven hiring process because trust in anyone you don't know goes to zero.

> The pessimist in me says that that just kills inbound entirely and we go back to an almost purely connections/network-driven hiring process because trust in anyone you don't know goes to zero.

What is the alternative; interviewing anybody who happens to apply? That is not feasible. Companies need a way to sort out the wheat from the chaff; they'll take any signal they can get.

> What is the alternative; interviewing anybody who happens to apply? That is not feasible. Companies need a way to sort out the wheat from the chaff; they'll take any signal they can get.

Connection. Every field will be as connection-based as politics.

And what if you are introvert or have social anxiety and not good at connecting to people? A lot of such people are very smart and exactly the type of people you would want to hire for some positions. Or what if you have a disability that makes it hard for you to travel and socialize? How about working remotely?

You will have a much harder time finding a job than people who don't share these traits.

I agree that they need a way to get signal. But resumes have never been a great way to get it - they just happen to be so cheap and every other signal so expensive at top-of-funnel that they've become the norm.

It's a vicious cycle. There's too much volume so we can't interview everyone, we can't interview everyone so candidates are incentivized to be spammy, so there's even more volume and we even more can't interview everyone. It's essentially the same reason that the old boomer "oh just walk in the door and hand them a resume" approach doesn't work anymore - it's too easy and therefore too spammy.

I used to be part of the leadership of one company working on this problem, and just founded another very recently, so it's something I've thought a lot about. I really think the thrust of the problem is:

* Signal is expensive to get, because

* The volume per candidate is too high

...and that the solution is to centralize early screening so you can get good signal. Which is what we're doing. Remains to be seen if it works, obviously.

IMO human recruiters and staffing agencies have done far worse than what AI could do.

My inbox used to be filled with 50+ resumes humans manipulated themselves - reviewing them for authenticity was hard and time consuming.

We make no claims to make this better. We claim to help the job seeker.

AI will indeed resolve dilemmas on all sides imo.

An output from the current state-of-the-"art" hallucinating models will be even more devastating.

What is a problem in editing your cv and simply being honest?

> we go back to an almost purely connections/network-driven hiring process

What if you don’t have friends? What if you are interviewer/not very social? Etc.

Video chat and real conversations are the only way forward at this point. Text is no longer a reliable signal as to how qualified or articulate the candidate is.

There are real stories to tell in text (prior experience, projects, certifications, tangible skills), and that sets the stage for the real conversations. Then it’s all about technical questions that require technical answers, and will this person thrive in our team.

The "stories" are not articulated by the one who experienced them. It's not their words. It's the words of AI.

Video is the only way forward. Let them speak on behalf of their own experience.

Couldn’t agree more.

I’m in the job market and I have resisted turning to AI for help with my resume. I feel that while AI could absolutely improve upon it the result would not represent who I really am. As a result I am enjoying an extended vacation.

I'm not sure if it's real-time yet, but video can also be faked well enough to fool people. Might just be real conversations left.

Although even then, when the AI gets competent enough… if you can't tell, does it matter?

Works both ways of course; when it's good enough in that sense, nobody is hiring either.

We're headed in this direction because people want seamless & intelligent automation in every aspect of their life. It's about the actual "Hey siri, go do x for me" being a reality.

And it's about AI's ability to actually deliver on that.

It's not about gaming the system or making things up. We try to guide all of the prompts we use to only expand upon or assume details from user prompts. We make strict mention to not make anything up.

I understand AI hallucinates, and alignment is early. But I'm a believer in where we are headed with AI, and I believe there will be a ton of valuable productivity to be gained.

Regarding what the other commenter said... I actually agree... let's kill inbound because human recruiters have been the most egregious players in hiring. As a tech leader I had countless bogus resumes thrown my way, long before AI was a reality. AI imo will revolutionize the hiring experience for the better.

But your product is producing human-formatted content for human recruiters. It sounds like you're interested in increasing the noise until there's no signal, and then hoping someone finds a way to bring the signal back.

We're interested in aiding the job seeker only – gain a fresh place to start with minimal effort. I can't say this enables or inhibits any sort of noise; other than intending to amplify genuine job seekers; again we try to be consistent in our prompts.

In other engagements I'm involved in, not through this product, we are seeking to enable employers. I do believe AI will help reduce noise; my argument is humans have been far worse in creating it than AI could be, and far less productive in filtering it than AI will be.

"sneek peak" should be "sneak peek"

At least we know you didn't use ai for the copywriting :)

thanks

We shouldn't be writing résumés. We should be writing academic- or engineering-style six pagers where the target audience consists of your peers (and not anyone under the org's HR umbrella).

This is why I think cover letters are maligned. They just shouldn't be mandatory. The best résumé is just a supercharged cover letter written by someone who actually gives a shit about what they're writing.

My 5-person team got almost 1000 applications for a job opening recently. We'd need AI summaries to have even a chance of sifting through all those 6-pagers, in which case we're better off if candidates just submitted a resume that was loosely targeted at our job opening.

You're already not looking at 1000 résumés.

I agree... But the reality is that what you suggest requires a significant shift in thinking and practice from a meaningful proportion of the very people you suggest we shouldn't be targeting. And they, in my experience, aren't exactly the kinds of people that are about challenging the status quo, rocking the boat, or moving beyond the received wisdom of their profession.

interesting idea, could you say more?

I think writing is a great idea, it was something I did earlier in my career. It showcased some interesting experiences of mine (matching algorithms deployed in the real world, porting neural nets to coreml in 2017, making advanced SSD configurations on kubernetes, etc)... these all gave me a lot of visibility and really only limited insight into my experience. Still, they enabled quite a bit career wise and plenty of conversation.

The best platform turned out to be Clubhouse... where I could talk the talk with like minded individuals... discussing preemptible VM hacking with Kelsey Hightower was memorable. But that platform sadly dies a slow death.

I think the cover letter is where experienced individuals have a chance to write the no-bs, all-out, explanation of why they are fit for the job. We choose not to offer this as a feature.

The resume is an anchor for interview conversations, and filtering in recruiting. I agree that it is all unideal actually but regardless is a part of the process. It can be a barrier for job seekers, and we wanted to break that barrier down.

I think AI is set to make the job seeking and hiring experiences better fwiw.

Reply on news.ycombinator.com