awb

10 ms·

1.
by awb 1mo ago
The demand curve for speed and intelligence seems pretty steep to me. If you look at the hiring marketplace, being just marginally better than your peers can be very lucrative. If you’re competing on speed or capability as a company (or as
2.
by awb 5mo ago
Employees can quit too. In both scenarios, it really depends on how many alternative options for income you have available.
3.
by awb 5mo ago
When you’re paying someone’s bill/salary, it changes the dynamic. Just like someone in customer service might act differently with coworkers vs customers.
4.
by awb 10mo ago
Exactly. It’s the concept of a management chart as an inverted pyramid with each layer holding up and supporting the layer above them. If you imagine a promotion as working your way down the corporate pyramid, then it’s easier to see how
5.
by awb 11mo ago
Some hypothesize that flashbacks might be the brain searching for relevant useful memories, or hallucinating if it can’t find any. Or, perhaps emotions or physical issues cause your brain to function differently and it’s not an adaptive tra
6.
by awb 11mo ago
But how do you measure intelligence or problem solving without language? It seems like an unavoidable and non-trivial parameter.
7.
by awb 11mo ago
The difference is the point of sale. With VS Code, you purchase your AI compute elsewhere (Anthropic, OpenAI, etc.), and then use it through the free VS Code interface. With Warp, you purchase your AI compute through Warp (who then pays Ant
8.
by awb 11mo ago
> On the Build plan, you pay for what you use and credits roll over month to month. Here’s where I got it from, but I see how it’s ambiguous. “You pay for what you use” sounds a bit like the BYOK (bring your own key) “add-on credits” pri
9.
by awb 11mo ago
> Pricing model for a terminal. What a time to be alive. You’re really paying for AI compute, not the terminal.
10.
by awb 11mo ago
Their old Pro plan at $15/mo (paid annually) had 2,500/mo AI requests per month, use it or lose it. The new Build plan at $20/mo has 1,500 AI requests, but they roll over. (Edit: apparently they don’t) > No bones about it:
11.
Investigating How Prompt Politeness Affects LLM Accuracy(arxiv.org)
4 points by awb 11mo ago|2 comments
12.
by awb 11mo ago
A few issues: * The casino takes a rake, so you lose money every hand, but you only win when the fish bets and loses. You’re also expected to tip the dealer * Everything is on camera and dealers remember players, so there will be a lot of w
13.
by awb 11mo ago
> Because they would be smart enough to know the societal damage they caused by this revelation and they did it anyways... That seems like a human-centric perspective. Maybe they’re a cooperative, altruistic society with an innate desire
14.
by awb 11mo ago
Hopefully there’s a similar “don’t make me think” mantra that comes to AI product design. I like the trend where the agent decides what models, tooling and thought process to use. That seems to me far more powerful than asking users to crea
15.
by awb 1y ago
For reference, here are the current guidelines: https://news.ycombinator.com/newsguidelines.html As a long-time user I’ve seen the most change around “What to Submit”: > Off-Topic: Most stories about politics > If the
16.
by awb 1y ago
Judging by the lack of responses and my own experience: no. Most subagent examples are vague or simplistic.
17.
by awb 1y ago
Fun fact: It would take ~278 hours to take a look at 1M websites for 1s each.
18.
by awb 1y ago
It would be interesting to analyze this dataset in terms of colors, layout, features, fonts, photos, etc. to be able to statistically measure the uniqueness or creativity of a given web design.
19.
by awb 1y ago
Interesting read and thanks for sharing. Two observations: 1. Natural language appears to be to be the starting point of any endeavor. 2. > It may be illuminating to try to imagine what would have happened if, right from the start our na
20.
by awb 1y ago
Hyper-niche products come with some inherent risk that it’s not always profitable to maintain or develop them long-term. With a mass market product leader you’re sacrificing a bit of customization for long-term stability.
21.
by awb 1y ago
PageRank worked well for Google for a long time. This sounds like an adaptation of that that’s interesting to consider.
22.
by awb 1y ago
Luxury goods and staple goods have distinct optimizations, both viable for generating profits and economic utility. A high end soup and an affordable soup might be serving two different markets.
23.
by awb 1y ago
IMO, it depends. I’ve worked with hundreds of startups over the last 20 years. The most significant trend I’ve seen is cultural cohesion. Some startups have a young, energetic culture. Some have an experienced, veteran culture. Some have a
24.
by awb 1y ago
As an engineer, if you’ve had a skilled Project Manager you’ve receive detailed, thoughtful “prompts” on how to complete a project that leads to a shared understanding. But with unskilled Project Managers you might receive a vague “prompt”
25.
by awb 1y ago
I’m not debating the potential accuracy of anyone’s interpretation, just commenting on the degree of simplicity of those interpretations. It’s a Razor, not a law, so it isn’t going to be 100% accuracy or predictive. But to suggest that the
26.
by awb 1y ago
Actually, Occam’s razor would suggest that their words be taken at face value. A hidden agenda is not the simplest explanation.
27.
by awb 1y ago
Many years ago there were guides on how to write user stories: “As a [role], I want to be able to do [task] so I can achieve [objective]”, because it was useful to teach high-level thinkers how to communicate requirements with less ambiguit
28.
by awb 1y ago
Your native language is now a programming language.
29.
by awb 1y ago
We’ve seen tech completely eliminate jobs like phone switch operators and lamp lighters. And it’s decimated other professions like manual agriculture, assembly line jobs, etc. It seems like people are debating whether the impact of AI on co
30.
by awb 1y ago
The best exercise is the one that you’ll do. In the vein of atomic habits, reducing a task to its minimal effective dose is great for starting new habits. For people that sit all day, a 1min exercise snack makes a lot of sense. After that b

More ›