The largest problem with a gpt future imo would be bot networks creating manufactured outrage. these will shift the decisions of policymakers towards those who controls these networks ie big corp.
1,838 karma · joined June 10, 2020
The largest problem with a gpt future imo would be bot networks creating manufactured outrage. these will shift the decisions of policymakers towards those who controls these networks ie big corp.
I was working at a Big Startup, on ML side of things. All the senior engineers in my team left and me and a few others became the senior-most engineers (with <1 YOE). The models we built were contributing 100M$+ to the top line of the business.
My then manager tasked to do some data analysis to determine whether building a "new nonsensical ML Model" was the right way forward. I like a diligent bee crunched the data and came to the conclusion that "we shouldn't build the new model".
I showed the results to my manager who laughed and told me to do it again. It's also important to note here that my team lead and old manager had left the company. This person was completely new and a senior manager.
I in my naive mind (and oblivious to politics) thought he's telling me to say that the results "do make sense". and that's what I said. We even presented the results to product people who gave us go ahead. Since we were doing such pivotal work, even the company's CTO was involved and tracking our progress.
The next few months were one of the most painful times of my life. Immediately after the meeting, we were tasked to actually build those models. I knew that they wouldn't bring any results, everyone else in my team knew, even PMs knew. Everyone except the manager.
I led a few interns and we built the damn models and got the expected result which is 0% improvement.
What happened next taught me a lot about life. Their was a meeting (which I wasn't a part of) where everyone senior who had their ass on this project came together. They invented a "new metric of success" and asked us to train ML models optimizing for that. We trained the models and obviously since the optimization target changed, we saw improvement.
Then we created charts according to that metric which were obviously good. the PMs showed slides to Upper management. and all was well again.
The best part of the new metric was that the actual results will only be visible years down the line. So we called it a success regardless. I switched my job soon after that.
I learnt from this experience to just say "NO" firmly on doing things I'm not comfortable with. and always keeping a text trail. The manager was kind of an oddball so I never fully got whether he actually knew this from the start or was just plain incompetant.
LLM is the swiss army knife for everyone not just developers.
One intern was told to deploy the app and I guess he made a mistake and updated k8 configs which deleted all the pods. Tell me how code reviews are supposed to catch it.
If you think "why not restrict config actions" etc, it's because the dev can go there and tweak those actions too. And each step added means less velocity of development. It's a game of cat and mouse. With isolation, you can limit the blast radius and loss of revenue.
The home page says that everything is private etc etc. Whatsapp also made similar claims before they were acquired, then suddenly these things didn't matter.
If this gets critical mass, what's stopping them from selling out. Bonus if they "accidently" started tracking users between update v4.2 and v5.2. Also, the TOS magically changed overnight.
The sad part is if they didn't sell out, they'll be buried under lawsuits and eventually banned. Just like KATorrents, yt-dl, Internet Archive, Dread Pirate Roberts etc.
1. https://www.businessinsider.com/suspected-kickass-torrent-ar...
Since the equity is basically worthless in this deal, I'd suggest going the "Mr Wonderful" way. Give equity + regular dividends until the (amount x 2~4)$ are paid.
This can be done for search as well which makes sure at least 4 players stay in the market at all times. Why isn't US govt doing this?
1. https://www.indiatimes.com/worth/news/upi-apps-may-soon-have...
Also, Can anyone tell why don't more people fly to other countries for these procedures? Paying something like 6k$ for a crown is pure extortion. In India, a decent one (with quality equivalent to US) costs around 60$. That a difference of 100x.
The hiring needs to be direct or else the "agency" will eat most of the money and give you 3000$/yr "senior" devs further perpetuating the stereotypes.
With 235k, you can make an incredible 5 people team + cloud costs for an year.
You can browse startup boards and generally you'll get 0.1% to 2% equity of the company for being a "founding engineer".
and then author continued this in his other post
" That person is really a founding engineer. Such a person is a special breed. Like an engineer on a larger team, the founding engineer needs to be able to execute and build product, quickly. They need to be aware of technical debt and when assuming it makes sense. They need to write code that works, is testable and performant. They need to evaluate whether to build or buy or download from Github. They’ll be performing devops tasks as well, such as monitoring the application, troubleshooting performance issues, and making sure deployments and rollbacks happen cleanly.
In addition to all of that, the founding engineer must communicate with non technical folks, both within the organization (including the CEO) and outside of the company. The founding engineer may or may not be doing tier 1 customer support, but will be doing tier 2 support. They also have to be willing and able to drive decisions, especially around technical issues. They need to be able to understand and give feedback on the business strategy at some level, as that will determine technical decisions. The founding engineer also needs to able and willing to take market feedback. This may be filtered through others, and any response should be discussed with the other founders, but market feedback and quick iteration is a strength of a startup. The founding engineer may be doing some product management as well, depending on the other team members’ strengths. They’ll be the de-facto UX designer. They also will be interacting with vendors and contractors. They may review budgets and will hire technical team members. It’s possible they’ll interact with investors if funding is sought.
For all the additional responsibilities, the founding engineer should be compensated, well, like a founder. With not much money and gobs of equity. (Not much money is obviously relative, and depends on market value.) "
I can't imagine what sucker would do all the tasks of CTO + multiple engineers for no pay and 0.1% to 2% equity and title deflation and no board seat.
Essentially the author wants to find a moron who doesn't realize how bad the deal is.
The problem is that the people who make poor decisions like this for themselves, also make such decisions for the company.
1. At Meta, you had to use internal "approved" tooling to build you product. which is far worse than products in open market. BlueSky can put it all on serverless like AWS Lambda and call it a day.
2. If funds are an issue, then raise more. There's no way around it. you had to scale sometime. Even if you optimise well, you'll still burn a huge amount supporting free users.
3. Threads was built by 10 Engineers. A simple MVP shouldn't have a massive team behind it anyways.
I feel the founders are largely responsible for not having their priorities sorted out.
Nobody is going to move again. They had the chance of outrun threads but slept on it.
Whatever they release now, will never get initial users to get it off ground.
Do you also expect a Med graduate to see some patients in the hospital for free before giving them a job?
The problem isn't Sensible vision.
The problem is getting enough users to make your community interesting.
The trust that VICTIM will get justice keeps the society intact. Nobody cares what happens to the bad guy.
If someone took out all your money through fraud and gambled it away, what would you choose:
A) Justice: Put them in Jail for long time.
B) Rehabilitation: Put them in medical care for an year till their gambling addiction is cured. Then they are free to integrate back into society.
In both cases, you get no money back. What would you prefer?
https://chrome.google.com/webstore/detail/coldmessageai-for-...
- Make a plan
- Execute it
- Execute Faster
- Even more Faster
If you get stuck in some parts such as Market Validation, then picking up the recommended book will be useful. It's easier to learn after failing.
I used to take months thinking about idea, market validation, half-decent codebase. Nowadays it's more of build in a day and try the best marketing channels according to the product. Then onto the next one. Improved "Turn Around Time" really helps in increasing number of iterations.
If it's loved(tolerated) in its worst possible self, then it has a chance.
Is it worth it to submit projects like this?
Otherwise I think all awards will automatically go to Nicky Case (ncase.me).