cloche

The guy had a failed startup and somehow leveraged that into the situation today.

You shouldn't always optimize for the long term. You should optimize for the reality of the environment that you're in. In large, public companies, that usually means looking ahead normally 1-2 quarters unless there are exceptional projects with leadership buy-in.

Unless you're working on the absolute cutting-edge, you don't need every member of the team to be at the top of their game. It's happened many times in sports that a team full of too many superstars underachieves.

Look at the President and the US government. This is society now.

It's attacking civilians that's controversial.

This sounds a lot like what some lazy people will say "I could be a doctor/astronaut/millionaire/movie star. I just don't want to"

Also, the one person listed in the Organization Members works at Google as a Developer Advocate.

Unless you're hiring for rare roles that require niche skills, it's unlikely that you'll get people that are in overlapping networks. If you do, you might be encountering the problem where Company A has a bunch of people that used to work at Company B. Now Company A is just an "old boys club" from Company B and is biased towards their old colleagues.

If you ask for single referrals, how do you combat the problem of people just recommended people based on friendships and not actually work quality?

Imagine you have a coin-flipping tournament with a large number of people - perhaps even billions. 2 people face off in rounds and the winner advances. By the end someone will have won all their rounds. Is this person skilled or lucky?

I don't understand why tech companies are so reluctant to go back to in-person interviews. This used to be the norm before COVID and it would solve most of these issues. It's also ironic that the authors are from Microsoft/Amazon and Meta which have very structured interviews that tend to be about just hitting the right keywords during the interview itself. None of these places even really care about what you've done previously, for better or worse.

He's basically said as much directly. I think it was in his book that he would like to get back to research at some point in the future. If he had things his way, AI would still not be public and would be used for advancing science only. Then there'd be no competitive interests pulling them into consumer products.

R&D is hit or miss. And a lot will miss. It's really no different from the VC model where 90% of their startups will fail but the 10% that succeed will make up for the rest.

Pointing out a few examples that didn't go anywhere is a meaningless argument. You need to look at it holistically.

Why would you want to ruin morale for no reason?

I track my time even though I don't have clients. I find it helps me say "ok this next block is for this task" and it helps me keep focus.

If you're being treated poorly and it's causing you stress, leave.

It's going to take time to earn trust from peers and your manager to start getting more meatier work. If you're early career, I think 2 years is a good guideline. Many places hiring someone for their second job will expect you to be leaving your first job around 2-ish years. 2 years gives you the chance to take on larger projects and see the results of your work and get feedback about things you've pushed to production.

You probably shouldn't stay at your first job more than 3 or 4 years. The second job change is the hardest. It's when you realize that different companies do and value things differently. Staying too long at your first job will make it tougher to adjust. It's also good to get exposure to new ways of working while you're still agile enough to soak it up.

If you've left more than 2 or 3 jobs within 2 years it starts to look like a red flag.

Never waste the opportunity a good crisis presents

You can only tell which part of the S-Curve you're on in retrospect. It's not something you can tell while you're experiencing it. Both scenarios of AI maxing out or continuing to improve are both likely.

I think this is it. I'm a general software engineer and would like to switch to being in this field but something like this is just way over my head. It sounds like a great innovation but I'm lacking the domain knowledge to fully understand it. I've spent several months going through online courses to basically get to a Biology 101 understanding. Getting to the level of understanding something like this seems like it would be a multi-year effort and I don't really know what's the best way to proceed.

It's pretty easy to lose 5 pounds. You need a total caloric deficit of about 17500 kcal over a certain period of time. There are plenty of caloric counting apps that are a few dollars/per month. Why would this be worth $100?

You're saying though that it can be done in one week without eating any protein and somehow you'll maintain muscle mass from taking a supplement of acai berry. Most of that weight loss is going to be water weight and glycogen, not fat. It seems dubious. All weight loss plans basically end up being "eat less calories that your body uses". There's no way around that.

You say the meal plan is 1800 calories. If you have a petite woman doing this, her caloric outake may be around that. Consuming 1800 calories for her is not going to do anything.

> this is actually a great video overview of how LLM works and the tokenization part

Did you mean to link to the video? I would be interested.

I was expecting legit questions. These are just rude

That was a time when the latest arms technology could be obtained both by normal citizens and the government. Do you think that today an armed population could repel a determined federal government?

Most innovation is just connecting the works of others together with a touch of individual creativity.

Have you used AI lately? Yes, this was the case a year ago but I've been using Claude daily for the past month and the quality is on par with what a senior engineer would be producing - often times better as it's more thorough and can think through all logical edge cases much more reliably. I do review all the code it writes. There are some suggestions I have but I'd say overall, 99% of it is solid.

I get that many people are tired of too many AI posts and there's an influx of negative comments but I think this is genuinely amazing. Think about watching Star Trek in the 60s and seeing people talk to computers and them understanding and being able to communicate back. We're literally living in the future!

The sun will last forever (at least from our point of view).

To maintain the muscle you have, you only need about 1/3rd of your normal workouts. It can be retained with 1-2 workouts per week. I imagine the same would be for something you've learned. If you've already put in the effort to learn it, reviewing it ~1x per week is probably enough. During the accumulation phase though - whether it be muscle or learning a new skill - once a week is definitely not enough.

Even so, the article says it grew 8% YOY in the US. The best is to hope that this is an unstoppable trend so that even politicians won't be able to reverse it.

I couldn't agree more. At this stage of my career, if I were to go back to an early stage startup, I would only consider a startup where the founder(s) had successful experience as at least a frontline manager. That provides a high chance that they will understand what it takes to successfully manage a team and, hopefully, a larger organization.

I think we'll need to evolve code review. Rather than simple rules like every review must have 1 or 2 approvals, we should be moving towards dynamic rules for when a human needs to be in the loop. For example, an agent can output a risk score for how risky the change is. Teams can then configure to auto approve PRs for the risk level they are comfortable with. Any risk level above that goes to a human reviewer. The advantage with this system is that it will help avoid massive vibe-coded PRs. If devs want to get their PRs deployed quickly, it will encourage a positive feedback loop of good engineering practices like keeping changes small and focused - that doesn't go away in an AI world.

Perhaps a summary is sent out to the team at the end of the day for all auto-approved PRs so that developers can review them quickly to stay up to date on the codebase.

AFAIK Github doesn't support anything like this today (maybe through actions?) so your company may need to use a tool that evolves in this way.

I'm the same as the parent. I've done this. I've been in groups for 2+ years and attend regularly and am not able to get any close friends out of it. I just don't form the connections that other people do. There's just something about the way I interact that doesn't cause me to get close to people. It might be something I'm doing subconsciously and I just don't know what it is. It's not like people can't stand me or anything. There's many people I can interact with normally at work or whatnot. I just don't seem to "click" with people. If anyone knows a coach or course that helps to address this, I would be interested.