atniomn

LLMs can write very performant code, but they have to understand that it's explicitly important to you.

If possible capturing golden inputs/outputs for a block of code, and just asking the agent to optimize while ensuring those goldens pass can be quite effective.

There's a lot of performance-sensitive code where I work (i.e. market data processing) and LLMs have been very effective at reducing our p50, p99, etc...

This doesn't really account for new ways in which we are social. After my kid goes to bed, I typically get an hour or two to spend online with my friends on Discord. We mostly talk about work, fatherhood, etc...

I think it's way more healthy for me to be able to talk about life every day for an hour online, than meeting up once per week for two hours over beer. I still see them in person on a monthly basis.

Don't know why you're getting hammered here. There's a clear skill factor in piloting LLMs.

GTM engineering? Customer-facing roles? I work at a firm with around 700 employees. We have an enterprise relationship with Anthropic and there have to be a dozen Anthropic employees in our shared Slack channel.

You're just moving the problem around. Instead of OpenAI and Anthropic dumping proofs, you'd instead have someone (maybe an aspiring graduate student, maybe a hobbyist, maybe an established mathematician who has disdain for existing process) dump these proofs en masse personally.

Large enterprises are paying unsubsidized metered rates. I work at a firm with around 700 employees. The p90 engineer is spending $2k/wk.

Our business-side employees have only begun using agents, and their spending accelerating. They're claiming their ROI is massively positive and have an enormous willing to spend.

Prices will probably go down because of the emerging price war between Anthropic and OpenAI, but there are many firms willing to spend.

I expect the next Anthropic release to finally reduce the prevalence of Claudish

I'm 32. In my senior year of high school, I woke up at 5:30 AM, my first class (zero hour) was at 6:30 AM, regular school started at 7:35 AM. school let out at 2:30 PM. I had extracurriculars until 6:00 PM most days. Go home, eat dinner while studying, go to bed at 11:00 PM. I took 8 AP classes and had straight A's. 5 on every test. strong ACT.

I attended a prestigious land grant university, like Michigan, and I was totally prepared and performed very strongly right out of the gate. I was competitive with elite students who had attended the most prestigious high schools both in the United States and internationally.

Today, I'm an extremely well compensated engineer. My household income is top 1%. My career is gratifying. I have a rich social life and a loving wife and child. If your children are operating at this caliber, and they're able to sustain this level of effort, they're going to go places, no matter what they chose to do. You are rewarded for this effort.

Splitting up Intel's fabrication and design businesses would it make similar to AMD, the stock which is up 45x since 2015.

Intel's fabrication business needs more volume, so it can increase R&D spending and capital expenditure, it can get this volume as a merchant fab.

Intel's design business doesn't need to be held back by Intel's fabrication delays. I imagine designers at Intel would prefer to compete with AMD on the same playing field--TSMC's latest node.