riku_iki

16 ms·

271.
by riku_iki 1y ago
> Because then the bigtechs will buy them all and everyone else gets nothing. and whats wrong about this? Humans funneled into area with higher added value.
272.
by riku_iki 1y ago
Why would they pay so much money for Chrome? They could just fork chromium for free? 34B is just for web address?
273.
by riku_iki 1y ago
> Polymarket betters are not impressed. Based upon the market odds, OpenAI had a 35% chance to have the best model (at year end) who will decide the winner to resolve bets?
274.
by riku_iki 1y ago
I think that 1.5m bonus is likely stocks with 500B valuation. There are other rumors they want outsiders to be allowed to buy stocks with 500B valuation.
275.
by riku_iki 1y ago
Uber had limited underpowered competition, so they could win starvation game. OpenAI competes with google, who can drop 50B/y into AI hype for very long time.
276.
by riku_iki 1y ago
OpenAI has clear and strange definition of AGI in contract with MSFT: it should produce 100B economic impact.
277.
by riku_iki 1y ago
> I wear thin shoes with no padding (barefoot style) so I couldn't even if I wanted to. if your form is wrong, it totally could result of extra impact on knee. Its better to run with padding than do not run at all. > It could be
278.
by riku_iki 1y ago
what would be reversing biological age from regular running and eating chicken breast + veggies with olive oil.
279.
by riku_iki 1y ago
Its unclear if Typescript has comparable backend ecosystem. For JVM you can find reliable, somehow well documented, widely used library for most of backend stories: driver for that database, SOAP-XML lib with support of niche security proto
280.
by riku_iki 1y ago
It could be because this is the easier way to build up resume and be competitive on the market.
281.
by riku_iki 1y ago
> Given their current product offerings, I really don't see a way they could ever justify a $300B valuation unless they get everyone on the planet to subscribe to their $200/month plan. the strongest opportunity is to compete w
282.
by riku_iki 1y ago
> American corporations reward performance over seniority more like loyalty to upper management which correlates with seniority (spent many years in the same company/group).
283.
by riku_iki 1y ago
there is also whole rdf/owl/semantic web ecosystem, with tons of companies/engines/datasets, mostly small.
284.
by riku_iki 1y ago
But such system exists, mostly implemented in prolog.
285.
by riku_iki 1y ago
> However, it does not appear to have been peer reviewed. my observation is that peer reviewers never try to reproduce results or do basic code audit to check that there is no data leak for example to training dataset.
286.
by riku_iki 1y ago
its survival bias, they were moving billions of years before, objects with unstable orbits traveled somewhere else, and stable objects formed stable planets.
287.
by riku_iki 1y ago
> I think self hosting is in a similar spot right now. The apps exist (many are extremely nice!), the software exists, but the seamless, aesthetically pleasing and ergonomic experience does not. It's a pain in the ass to set up self
288.
by riku_iki 1y ago
I would argue is that some portion of ecosystem should be readable by humans. In programming engineering we already have this: there is human readable high-level code, and there is assembler and lots of auto-generated code. In proof system
289.
by riku_iki 1y ago
> I can see how one could be tricked into the idea that success mainly comes from a good idea and not a good execution. in your situation, if you would have good idea which solves real problem, you would see at least some people come to
290.
by riku_iki 1y ago
language codebase, they have some tests there for example, using which model can see what are the correct typical implementation details. > And documentation doesn't have the volume LLMs need why do you think so? It looks like LLM h
291.
by riku_iki 1y ago
There were so many embarrassing topics about this, that openai for sure added it to training dataset with high priority
292.
by riku_iki 1y ago
LLM can give many answers using absorbed docs and codebase. Rest still could be asked/answered on SO or github.
293.
by riku_iki 1y ago
Yes, makes sense, thank you.
294.
by riku_iki 1y ago
> if that's you feel there's probably not much I can say to change your mind you just brought several corp statements which are not grounded into any evidence, and could be not true, so you didn't say that much so far. &g
295.
by riku_iki 1y ago
> such announcements should wait at least a week after the closing ceremony it would raise more concerns that corps leaked questions/answers to training data and finetuned specialized models during this time.
296.
by riku_iki 1y ago
there is also huge realm of private/commercial data which is not absorbed by LLMs yet. I think there are way more private/commercial data than public data.
297.
by riku_iki 1y ago
because competitive coding is narrow well described domain(limited number of concepts: lists, trees, etc) with high volume of data available for training, and easy way to setup RL feeback loop, so models can improve well in this domain, whi
298.
by riku_iki 1y ago
> The issue is that trust is very hard to build and very easy to lose I think its opposite: general public blindly trusts all kind of hyped stuff, its a very few hyper-skeptical who are some fraction of percent of population.
299.
by riku_iki 1y ago
because per corps statements, AI are now top 0.1% of PhD in math, coding, physics, law, medicine etc, yet, when I try it myself for my work it makes stupid mistakes, so I have suspicion that corp very pushy on manipulating metrics/benc
300.
by riku_iki 1y ago
there is standard approach: you setup raid, and FS will restore as much as possible and likely everything. Adding extra complexity to cover some edge cases maybe is too overkill.

More ›