theptip

23 ms·

481.
by theptip 1y ago
Absolutely, my experience too. I think the bleeding edge models are very good at “idea infill”. Depending on your subject matter you might only need an idea or two per 100loc generated. So much of what I used to do turns out to be grunt wor
482.
by theptip 1y ago
> At that point, a shared ledger implemented with traditional databases / protocols would be faster, easier, and more transparent. What open-source shared ledger would you suggest is a better fit?
483.
by theptip 1y ago
> That's a bit pedantic It's a pedantic reply to a pedantic point :) > If I'm prompting a LLM, what's the bridge from prompt engineering to software engineering? A sibling also made this point, but I don't fol
484.
by theptip 1y ago
I don't follow; you can read the code that your LLM produces as well. It's way easier to drill down in this way than the bytecode/assembly vs. high-level language divide.
485.
by theptip 1y ago
> and it never will be The only thing I’m certain of is that you’re highly overconfident. I’m sure plenty of assembly gurus said the same of the first compilers. > because it's not an abstraction This just seems like a category e
486.
by theptip 1y ago
Just beware the “real programmers hand-write assembly” fallacy. It was said that compilers would produce a generation of programmers unable to understand the workings of their programs. In some sense, this is true! But, almost nobody thinks
487.
by theptip 1y ago
Doesn’t actually seem like double-dipping. Users are paying for a service that was costed 5-10 years ago based on human web traffic. Now AI crawlers are a new source of huge traffic volume and CF is figuring out how to cover costs or profit
488.
by theptip 1y ago
I would endorse this position at very least directionally. The job is going to move up the abstraction stack; one person will do what a 5-person startup did, which means a founder must own more product and sales stuff. If you like working i
489.
by theptip 1y ago
> and therefore aggressively sought to showcase my leadership within the first few weeks I feel this is as close to “Management 101” as it gets, or even “employee-ing 101”, as it applies to technical roles too. Don’t do it! Just write do
490.
by theptip 1y ago
I feel there are a lot of workflows that are actually quite suited to Claude Code - for example here, running the MCP tool to iterate on a flight plan, and updating a .md file (bonus points - auto sync’s to your iCloud Drive and thence into
491.
by theptip 1y ago
Let me tell you about this app called “X”, formerly known as “Twitter”. People use it to post their opinions about products, among other things.
492.
by theptip 1y ago
I think you’re being way too cynical. The first sentence talks about risks: > When AI can interact with web pages, it creates meaningful value, but also opens up new risks And the majority of the copy in the page is talking about risks a
493.
by theptip 1y ago
Yeah good point. I think the point about cost is the real crux here, regardless of working temp. (CSP has used molten salt but it is quite expensive to store and maintain.)
494.
by theptip 1y ago
Water will be limited to 100C, and tanks are going to be way more expensive to maintain. You don’t get much long-term storage with your black paint as it’s just going to heat the surface. The point here is to heat up a big enough pile of di
495.
by theptip 1y ago
Yeah, TFA clearly explains their point. They mean RAG=vector search, and contrast this with tool calling (eg Grep).
496.
Building ultra cheap energy storage for solar PV(austinvernon.substack.com)
99 points by theptip 1y ago|94 comments
497.
by theptip 1y ago
Good context, thanks. If it’s likely that the Foundry will be sold off, do you view the extra 5% as a short-term profit and not really a geopolitical/strategic protection? Seems the Foundry is the bit that USG really is trying to backs
498.
by theptip 1y ago
> Finally it occurred to me to put context where it was needed - directly in the test files. Probably CLAUDE.md is a better place? > Too much context Claude’s Sub-agents[1] seems to be a promising way of getting around this, though I
499.
by theptip 1y ago
> the government made an $8.9 billion investment in Intel common stock, purchasing 433.3 million shares at a price of $20.47 per share, giving it a 10% stake in the company > Of the total, $5.7 billion of the government funds will com
500.
by theptip 1y ago
There is far more cultural diversity between the EU countries than states in the US - it’s not close.
501.
by theptip 1y ago
OpenAI is the fastest growing consumer tech company of all time. The big guys have to move at what for them is breakneck speed, just to have a chance at retaining their market share. They will be swept away if they sit and wait a couple yea
502.
by theptip 1y ago
I agree with this as one reason Apple hasn’t rolled out Apple Intelligence. It’s a fair point. However, I think it was a strategic blunder to go announce a full guns-blazing release that solves all of these problems, before taking the easie
503.
by theptip 1y ago
Agreed. I think the latency wins could be meaningful for many of the obvious phone-centric usecases (Siri, app-/browser-enrichment, etc.) but the capability gap is definitely a hindrance. Privacy is already a key differentiating featur
504.
by theptip 1y ago
As an aside, anyone care to speculate on the “secret seat configuration”? I guess maybe it’s a recliner with feet pointing to the outside (maybe just two seats per row)? That’s the only new configuration I can imagine that would require res
505.
by theptip 1y ago
I guess I’m using the product wrong if I find them useful?
506.
by theptip 1y ago
Yeah exactly. The MacBook Pro is by far the most capable consumer device for local LLM. A beefed up NPU could provide a big edge here. More speculatively, Apple is also one of the few companies positioned to market an ASIC for a specific tr
507.
by theptip 1y ago
> Google has world-class research teams that have produced unbelievable models, without any plan at all on how those make it into their products beyond forcing a chat window into Google Drive. NotebookLM is a genuinely novel AI-first pro
508.
Reflections on Neuralese(lesswrong.com)
1 points by theptip 1y ago|0 comments
509.
by theptip 1y ago
But for the case of Perplexity-User, presumably the user is in the loop to provide their attestation. This case (“go research this subject for me”) is the grey area here. It’s not the same as simple scraping or search indexing, it’s a new a
510.
by theptip 1y ago
It’s fair, because they earned it by building the best offering on the market. Fairness doesn’t mean everyone gets funded regardless of their quality.

More ›