overfeed

15 ms·

271.
by overfeed 4mo ago
Remember the context - this is while solving a whiteboard problem. Its bad in the same way asking candidate what their birthstone is - because any answer offer little or no signal about the odds of a candidates success at the company. I&#x2
272.
by overfeed 4mo ago
I wish you had addressed at least one of arguments in good faith before jumping to insults and countering a strawman argument I didn't make - I never claimed their will be no use for more capable models. You do your AI-maximalism, and
273.
by overfeed 4mo ago
> The most impressive think that stuck with me is that humans are incredibly efficient, from an energy perspective, in anything we do, compared to machines. Humans are efficient, but not across the board. Trivial counterexample: walking
274.
by overfeed 4mo ago
> On the other hand, even if Apple's AI were 6 - 9 months or a generation behind, Do you mean Google's AI with Apple wrappers? Apple's in-house AI is further behind Google, amd very far from the frontier according to yo
275.
by overfeed 4mo ago
Not only that, but such sponsorships would also generate new, high-quality end-to-end training data from targeted projects, including work-in-progress edits and prompts which are not present in scraped git histories.
276.
by overfeed 4mo ago
I see your pooint and disagree with it.
277.
by overfeed 4mo ago
> A 2026 4B beats 2024 4B, but both are far behind the contemporary frontier. The thing about engineering is you don't just use the biggest bolt on the market on every bridge. > In 2024, a "good" model is one that can b
278.
by overfeed 4mo ago
> ...the research progresses before the inevitable nationalization of the frontier. Hacker News has been telling me America beats China at "innovation" because of the "freedoms" - especially frew enterprise. I wonder
279.
by overfeed 4mo ago
> Wrong, mostly. > Model capability is a function of model size Model effectiveness has improved across model sizes. You really should try the latest flash variants more. They have become my default for most tasks except for gnarly hi
280.
by overfeed 4mo ago
> Not where I used to work It can't have been at any of the big 4, because partners aren't skipping 4+ org-chart layers to look at draft documents written by early-career associates. I have no experience with body shops - if th
281.
by overfeed 4mo ago
Have you seen the job market lately? You call the response "cagey and evasive", but that is for an objectively a bad interview question, one wrung below "How many years experience do you have prompting Anthropic Opus? We are
282.
by overfeed 4mo ago
The amount of pro-AI AI bots on chat forums is insane.
283.
by overfeed 4mo ago
> These six-figure reports are produced by underpaid kids in their twenties working 18 hours a day. That's accurate, for the first draft. Similar to big legal firms - subsequent versions are signed-off and passed up (and if revision
284.
by overfeed 4mo ago
Some laboror has to cask it first.
285.
by overfeed 4mo ago
> So, are you saying that the employees were exploited in some way? Google ads "extracted" value from traditional advertising in newspapers and magazines, so the "exploitation" (or efficiency gains, if you're cha
286.
by overfeed 4mo ago
I doubt considerations are based on need. Filestash is cool, but probably isn't the marquee marketing opportunity they are looking for; it jas to be a household name they can name-drop or place a logo on a marketing page and get instan
287.
by overfeed 4mo ago
> This is super common with startups and is usually called an orderly shutdown Perhaps now, but during the Zero Interest Rate era, the received wisdom was founders ought to keep going until there bank account was empty, in the hope that
288.
by overfeed 4mo ago
> I ended up just filing a few issues and moved on to other things. This is the most valuable contribution you had time for, hopefully with a minimum-viable bug reproduction. Drive-by patches/PRs are usually a net-negative because
289.
by overfeed 4mo ago
Chinese AI models also share a positive trait: they offer more bang for the buck.
290.
by overfeed 4mo ago
ITT: "What is "coordination", and why is it bullshit?" The lack of humility among tech folk is astounding. Why dont you ask yourself why the 10x'ers/doers/high-impact people aren't setting up their 3-
291.
by overfeed 4mo ago
> But this is solvable That's the crux of my point; Apple could have solved this on day zero if they had a consumer-centered threat-model and/or considered user data to be a liability rather than a hook for service subscription
292.
by overfeed 4mo ago
Apple's concern is at the intersection of DMA and privacy. Apple is worried that other parties having the same level of data access that Apple has today would create privacy issues. This is because Apple's current privacy postur
293.
by overfeed 4mo ago
> The founders retaining voting control has worked out fine for Alphabet shareholders. Only after the cofounders brought in veteran "adult supervision" who offered guided the company with a steady hand, while Larry and Sergei w
294.
by overfeed 4mo ago
It's not "fake" - it's cultural differences where what is intended to come across as polite by Americans[1] can be seen as insincere by people from elsewhere. On the flip side, Americans often view foreign behavior that&
295.
by overfeed 4mo ago
Big labs ripped videos off YouTube without caring about the ToS, and grabbed as much published literature they could get their hands on, regardless of legality (Books3, The Pile). The goal of "democratizing human knowledge" by way
296.
by overfeed 4mo ago
> Esp check the Hallucination rate for Deepseek - it's not good. For strongly-typed coding tasks - and I imagine other tasks that have cheap validity checks: agentic harnesses and thinking tokens are an effective foil against halluc
297.
by overfeed 4mo ago
> Or pay a guard a fair wage and comp them the $0.20 or whatever for each bullet. I think $0.20 per bullet is far too little, considering the medical expenses the guard will face when getting the bullets removed after they are shot for c
298.
by overfeed 4mo ago
> That would fit quite well with the general theme of prioritizing the interests of investors over all else. It's not a "general theme", it's right there in the name of the economic system.
299.
by overfeed 4mo ago
> but nobody uses AMD cards AMD is selling every MI card it makes, and the market wants more of them.
300.
by overfeed 4mo ago
> It wasn't just that crypto was an obvious grift Was it universally obvious? Hindaight is 20/20 There were many block-chain startups funded, and even FAANG got caught up in the hype. FWIW, I was a crypto sceptic, but I had man

More ›