lumost

15 ms·

271.
by lumost 1y ago
The big problem with AI in back-office automation is that it will randomly decide to do something different than it had been doing. Meaning that it could be happily crunching numbers accurately in your development and launch experience, t
272.
by lumost 1y ago
We don't have evidence one way or the other. But from the public statements the idea that they lose roughly their revenue seems constant over time. It's possible that that is simply a psychological barrier for investors. Meaning t
273.
by lumost 1y ago
If OpenAI continues on their current revenue growth trajectory, they should be larger than AWS by 2027. Burning 2x revenue to grow that fast is not really a concern beyond your continued ability to attract financing. Given the trajectory of
274.
by lumost 1y ago
At least testably/symptomatically, I'm asthmatic as well - so it's surprising that the impact is so small. My wife gets it for 1-2 weeks whenever she comes down with it.
275.
by lumost 1y ago
It's my unfortunate experience, when I've had covid its a 6-12 hour affair that happens once every 12-24 months. My 3rd vaccine shot had me in bed for 3 days. Leading to continued vaccination being unsustainable. My wife has a sim
276.
by lumost 1y ago
It's tough to get people to want a vaccine which knocks you off your feet for 3 days and needs to be repeated every 6-12 months. I'm very bullish on mRNA vaccine technology - but it's potentially a poor fit for rapidly changi
277.
by lumost 1y ago
Hiring irrational players, or forcing rational people to act outside of their own self-interest is not a winning strategy either. There is nothing wrong with building stuff, or career development. There is also nothing wrong with experiment
278.
by lumost 1y ago
I suspect the common issue with small scale projects is that it's not atypical for the engineers involved to perform a joint optimization of "what will work well for this project", and "what will work well at my next pro
279.
by lumost 1y ago
Keep in mind, L6 senior engineer at Amazon maps to staff at most firms. L5 has a crazy wide band to accommodate everyone from folks with 2 years experience in-house to 10+ years of experience.
280.
by lumost 1y ago
Amazon pay had a 30% negative revision for most tenured staff this year. It’s unclear how it plays out with new hires - there are likely still a few very strong Covid era grants out there. It’s not bad pay, but most mid-caps should be compe
281.
by lumost 1y ago
The fundamental constraint the article alludes to is the powerlessness of consumer choice. You can’t make a better choice because you don’t have any better option. When there is a better option, you lack the tools to verify that the option
282.
by lumost 1y ago
Or that this system would fail to adapt in anyway to changes of circumstance. The adaptive intelligence of a live human is truly incredible. Even in cases where the weights are updatable, We watch AI make the same mistake thousands of times
283.
by lumost 1y ago
There needs to be roles for pure ICs. In deep tech orgs, it’s almost impossible to get work done unless you have a bench of folks with domain expertise, comfortable on your software, and in tune with and guiding product direction. Diluting
284.
by lumost 1y ago
lot's of columns are float valued, GPU tensor cores can be programmed to do many operations between different float/int valued vectors. Strings can also be processed in this manner as they are simply vectors of integers. NVidia pu
285.
by lumost 1y ago
It does make me wonder whether all of the investment in hot-loading of GPU infrastructure for LLM workloads is portable to databases. 30TB of GPU memory will be roughly 200 B200 cards or roughly 1200 per hour compared to the $240/hour
286.
by lumost 1y ago
I'd doubt that they intend to spend a full trillion. The spend commitment is such that no one else can plausibly outspend them. If you take the position that spend is correlated with outcome - then they are on a positive track to win.
287.
by lumost 1y ago
The doomerism on OpenAI finances is unfounded IMO. They will survive and be a large company at this point. The big question marks are on just how big, when, and the cost to get there. If they lost all financing tomorrow, they'd deploy
288.
by lumost 1y ago
This depends on the scale of company. A fully functional DR plan probably costs 10% of the infra spend + people time for operationalization. For most small/medium businesses its a waste to plan for a once per 3-10 year event. If you’re
289.
by lumost 1y ago
Aye- given the margins involved, is imagine you could get quite favorable insurance policies from NVDA on tier 0 facilities.
290.
by lumost 1y ago
How much does the capex model of a datacenter change when the goal is 100% utilization, with no care for node uptime beyond capex efficiency/hardware value mainenance? I wouldn't be surprised if the cost came down by at least one
291.
by lumost 1y ago
If I have a user facing upload button, why can't I simply have a webserver that receives the data and pushes it into s3 via multi-part upload. Something that can be written in a framework of your choice in 10 minutes with 0 setup? For
292.
by lumost 1y ago
I've observed massive back office pipelines using dozens of interconnected lambda, batching, streaming, distributed storage for ephemeral data and other rube Goldberg contraptions to build what was ultimately a cron job on a modest ser
293.
by lumost 1y ago
I do wonder how much nymbyism is influenced by most individuals carrying 5-10 years earnings in their home. Such a potential liability might make one awfully concerned about liabilities.
294.
by lumost 1y ago
To what extent is our understanding of this period limited by survivorship? Granite doesn't weather significantly, so we see lots of metalithic stone structures with ambiguous dating. The near east had significant changes in inland cli
295.
by lumost 1y ago
While assets like stocks do better when currency depreciates in the short term, unless customers have proportionately more currency the price has to fall in-line with the currency in the long run. Gold etc. track a basket of currencies more
296.
by lumost 1y ago
I'm very curious if anyone has tried to control for the natural hierarchies which form in Academia. e.g. A researcher who rises to the top of a funding collaboration will have a disproportionate number of citations due to their influen
297.
by lumost 1y ago
Exactly right! A corporations internal governance has only a loose correspondence to a market economy. It's quite common for individuals to be promoted who have actually been counter-productive to the stated market goals of the organiz
298.
by lumost 1y ago
At the time of Adam Smith’s wealth of nations, there was a schism between eastern and Western Europe on the role of politics in the economy. Eastern Europe was influenced by much stronger aristocracies that exercised stronger control of the
299.
by lumost 1y ago
This seems to be OpenAI's path to victory in the AI race. Buy up the supply chain of compute to the extent that no other competitor could possibly have access to the same compute. If they are able to shut out Google/X.AI from the
300.
by lumost 1y ago
This depends significantly on the real cost of the gpu. The cost of enterprise ai GPUs is likely 2 orders of magnitude lower than the current list price. These deals avoid the need to markdown the price of the hardware to gain volume. If I

More ›