lumost

31 ms·

361.
by lumost 1y ago
It's not so much a question about fairness, just that the employee and employer are playing different games. Acknowledging this and devising a compensation strategy which aligns incentives is important. Employers play an iterated game
362.
by lumost 1y ago
There is no mysticism in incentive structures. My point was rather that if you provide strictly below market compensation (as most startup equity is positioned these days). You are likely to get below average talent, or below average result
363.
by lumost 1y ago
For early/mid stage startups - this is an awful position to take. These orgs are heavily influenced by who they hire - what you pay defines your incentive structure. Does the world class engineer or business development lead just take
364.
by lumost 1y ago
It depends, there are certainly founders who come from sufficient money that the value of founding a unicorn is not material to their Net worth. However I’d also suggest that the concentration of tech in the last decade is also partly due t
365.
by lumost 1y ago
I'll wait until some more benchmarks are run in this case. Unlike traditional software, vetting a model architecture works better than alternatives is a time and compute intensive process. You really can't just download it and &qu
366.
by lumost 1y ago
Did it work? :) The architecture is very similar offset lstms which have been studied extensively. The main difference is the handover of the hidden state, which my naive mind would assume makes optimization substantially more difficult.
367.
by lumost 1y ago
I am extremely skeptical of a 27M parameter model being trained “from scratch” on 1000 datapoints. I am likewise incredulous of the lack of comparison with any other model which is trained “from scratch” using their data preparation. Instea
368.
by lumost 1y ago
I’ve managed a few 8 figure infrastructure bills. The exponential decrease in hardware costs combined with the exponential growth in software engineer salaries has meant that these bills become inconsequential in the long run. I was at one
369.
by lumost 1y ago
The percentages in the field are skewed, FAANG employ a vast number of engineers.
370.
by lumost 1y ago
once upon a time - engineers often had to concern themselves with datacenter bills, cloud bills, and eventually SaaS bills. We'll probably have 5-10 years of being concerned about AI bills before the AI expense is trivial compared to t
371.
by lumost 1y ago
There are a lot of venues through which collusion has become legal. Management consulting, private equity investments, conferences etc. The incentives to collude are powerful.
372.
by lumost 1y ago
If you are trying to treat the agent as equivalent to a full time senior engineer that works 10x as fast… then you will be sorely disappointed. Right now the agents are roughly equivalent to a technically proficient intern that writes code
373.
by lumost 1y ago
The latest copilot pricing for Claude sonnet 4 is fantastic. I get 1500 agent sessions for $40/month
374.
by lumost 1y ago
My 2 cents, AI tools benefit from standard software best practices just like everyone else. I use Devcontainers in vscode so that the agent can't blow up my host machine, I use 1 git commit per AI "task", and have CI/CD
375.
by lumost 1y ago
I think we'll see a shift to BYO LLM in the future. I saw that overleaf offers an assistant feature and an associated subscription. But I already pay for ChatGPT and Github copilot, I'm not going to pay for a third assistant - par
376.
by lumost 1y ago
> Very wealthy people, politicians, and corporations are the beneficiary of the government largess. Is the proposal to cut large contracts or replace them with civil service? How do you reclaim these expenditures? My anecdotal observatio
377.
by lumost 1y ago
The issue here is that no one can agree on which portion of government spending is wasteful. It’s almost taken as a tautology that the government is wasteful. After 50 years of cost cutting and wage freezes, I tend to find it fanciful that
378.
by lumost 1y ago
The US can easily afford its obligations. The problem is that the tax base has shifted into alternative forms of income such as dividends and cap gains which are taxed at lower rates. It’s a policy choice that these returns are not taxed at
379.
by lumost 1y ago
The free tiers might be tough to sustain, but it’s hard to imagine that they are that problematic for OpenAI et al. GPUs will become cheaper, and smaller/faster models will reach the same level of capability.
380.
by lumost 1y ago
Anecdotally, usage rises precipitously when you are building a system from scratch with unlimited ai access.
381.
by lumost 1y ago
Im sure it would alert me to the trend, just as the scale does :) the fundamental problem is whether health is being exchanged for another good “stable income.” No calorie counter will stop a ramen quest after 90 hours of work. Unfortunatel
382.
by lumost 1y ago
The ordering out tends to correlate with someone not having the energy or discipline to make what they want themselves. That lack of energy leads to caving on food cravings and overeating. I recently switched from a major tech company to an
383.
by lumost 1y ago
You're being downvoted as many people who try to count calories fail to achieve meaningful results. From my own experience, weight gain follows a simple progression. 1. Expected high stress work day -> Coffee w/ food item in th
384.
by lumost 1y ago
I'd imagine there are multiple approaches to this e.g. shredding heads which break up soft material in front of the bot, or multi-robot configurations for snaking material. There might be an efficiency gain to be had in overall city pl
385.
by lumost 1y ago
Raising requirements in the US is prohibitive due to the lack of options individuals have when they cannot drive. A 16 year old often needs, or wants to work for an income. Owning a car is often a job requirement due to the distances invol
386.
by lumost 1y ago
As I recall, the majority of these paradoxes are resolved by the observation that you can't get a free measurement, for one system to measure another - both must interact. So either 1. You Perform some form of strong measurement which
387.
by lumost 1y ago
Low volatility is good for everyone engaged in long term asset management. Jane Street just found a way to make everyone else less money while making a small amount for themselves.
388.
by lumost 1y ago
Wouldn’t this be largely corrected by reference checks/letters in hiring? I’ve seen those becoming more frequent.
389.
by lumost 1y ago
The finite nature of btc, low transaction volume, and increasing cost of mining made deflation a given. The original designers simply did not solve this problem. BTC’s dominance in the crypto community suggests that this trait was advantage
390.
by lumost 1y ago
The vast majority of a I startups will fail for reasons other than model costs. If you crack your use case, model costs should fall exponentially.

More ›