lumost

8 ms·

151.
by lumost 7mo ago
I suspect the issue is the SDLC methodology of existing mature products. The "I can build it in a weekend" use case has gotten a massive boost as you can build something which "looks" real faster then ever. Mature teams
152.
by lumost 7mo ago
Or its a simple signifier that the author was human, and that a real person is trying to convince you of something. I've experimented with putting minor grammar mistakes into my work of the sort that would be frowned upon, but are not
153.
by lumost 7mo ago
Hedging against currency collapse is notoriously difficult. But a reasonable set of hedges are assets that will have value regardless of currency changeovers. Eg housing, land, gold etc.
154.
by lumost 8mo ago
Given Kuwaiti air force pilots would have dealt with Saudi/US/Iraqi F-15 operators, that seems highly unlikely.
155.
by lumost 8mo ago
The US is not acting rationally on Iran, this is on top of a history of acting irrationally with regards to past treaties. The game theorists would tell you that Iran should impose costs for this in one way or another, the main avenue of wh
156.
by lumost 8mo ago
Prices, and feelings about the economy are set at the margin. A 1% change in employment doesn't sound world ending on its own, but the relative power balance between labor and management pushes wages down across the economy.
157.
by lumost 8mo ago
Shahed style drones are quite inexpensive, the straight between key west and cuba is only 90 miles. To my knowledge, it is practically impossible to destroy distributed shahed launchers from an air campaign alone.
158.
by lumost 8mo ago
It's a good point, it would be a nice "upgrade story" to get the next generation model. At a fixed cost of ~$1000 per model, it wouldn't be a bad deal relative to current api costs.
159.
by lumost 8mo ago
The latest strategies of etching weights into silicon seem like they can be generalized. We currently design gpu/tpu caching on the basis that the weights change frequently - if the weights do not change at all, or change very slowly -
160.
by lumost 8mo ago
The Iran war is symptomatic of what's to come. If the US/other major powers feel unconstrained to wage semi-limited disruptive conflicts at will - then the world will be much less stable. Iran can close the straight of hormuz as r
161.
by lumost 8mo ago
It's more about the size of the task I try to do, it's quite possible to get opus4.6 to one shot a "good" 30k loc change with the right planing doc. I'm not confident I could get similar results handholding. I also
162.
by lumost 8mo ago
I’ve observed radically different workflows amongst senior candidates vs junior candidates when using an ai. A senior candidate will often build an extremely detailed plan for the agent - similar to how you would do a design for/with a
163.
by lumost 8mo ago
All limited exchanges end in annihilation, parties are required to tit for tat the exchange and it's really easy to run out of targets when you are using nukes - forcing an escalation. If Taiwan was invaded, the only military targets f
164.
by lumost 8mo ago
Why give munitions to an unreliable partner? why rely on an unreliable partner for munitions? the Japan/Australia/SK-US-EU/NATO alliance depends on the US being sensible and broadly aligned with the goals of the other countri
165.
by lumost 8mo ago
If China was to attack Taiwan, now would be the time. The current world order is at least in part based around the notion of the US (and allies) having the military capacity to fight any plausible combination of foes at all times. That this
166.
by lumost 8mo ago
Anthropic is running a similar marketing campaign as AWS/Devops tools which were trying to replace in-house IT. Pitch to the few that you can be 10/100x as productive and valuable on the hopes that they will push their organizatio
167.
by lumost 8mo ago
Didn't oracle take out real loans and spend real dollars based on this commitment?
168.
by lumost 8mo ago
Tesla and similar companies really make me wonder if we still live in a capitalist system. If wealth is sufficiently concentrated - the value of anything becomes tied to the whims of the few who can transact at that level. How a stock goes
169.
by lumost 8mo ago
Also makes you wonder how much of the user growth could just be distillation attempts from one model vendor to another.
170.
by lumost 8mo ago
Also - how can this be prevented? the AI labs can't seriously expect that each lab will filter LLM generated content from their training sets based on the source model. Leakage of AI behavior into public datasets is inevitable.
171.
by lumost 8mo ago
The question is how much further and faster can these tools evolve. Right now, you may not be faced with a valid choice to vibe code your own hubspot, but maybe some contractor firm will do it for you and sell an ultra low cost version. Wil
172.
by lumost 8mo ago
Curious if part of this was the overall decline in government compensation relative to the private sector. The president makes roughly what the typical SV engineer makes after 5 years in big tech or as a fresh grad from a top PhD program. M
173.
by lumost 8mo ago
It's quite plausible to me that the difference is inference configuration. This could be done through configurable depth, Moe experts, layers etc. Even beam decoding changes can make substantial performance changes. Train one large mod
174.
by lumost 8mo ago
This really points to a world where all services are too cheap to meter. The compute side of AI is a commodity, the usage of AI is a commodity, the model development of AI is a commodity. So far there is no evidence that a provider with hea
175.
by lumost 8mo ago
It's a good way for a grocer to minimize waste. When raw chicken gets close to sell by date, turn it into rotisserie chicken, when it doesn't sell - turn it into sandwiches and other products.
176.
by lumost 8mo ago
Personal agents disrupt OpenAI’s revenue plan. They had been planning to put ads in ChatGPT to make revenue. If users rapidly move to personal agents which are more resistant to ads, running on a blend of multiple models/compute provid
177.
by lumost 8mo ago
Anthropic's primary Capex partner is AMZN. AMZN is presently willing to drop 200 billion a year into Capex for compute to rent to anthropic and others. This 30 billion only needs to fund their rental rates - unlike openai and google wh
178.
by lumost 8mo ago
Does it really? Or does it become a panacea to make people feel like they’ve done something when they haven’t? Perhaps if these non-solutions didn’t exist to appease our fears then there would be more pressure for real solutions.
179.
by lumost 9mo ago
I don’t understand why most cloud backend designs seem to strive for maximizing the number of services used. My biggest gripe with this is async tasks where the app does numerous hijinks to avoid a 10 minute lambda processing timeout. Rathe
180.
by lumost 9mo ago
The history of software has been that once it becomes cheap enough for teams to flood the market with “existing product” + x feature for y users. The market consolidates around a leader who does all features for all customers. I’d bet that

More ›