dpflan

42 ms·

601.
by dpflan 4y ago
Sure, it will happen, but maybe it can be better regulated. But maybe it goes well with those ivermectin doses you also got prescribed over Zoom...
602.
by dpflan 4y ago
Alcohol has an extremely long history in humanity; synthesized chemicals less so (comparatively speaking). The alcohol industry is ingrained and well established and clearly wants to keep its influence and profits. Prohibition in the US in
603.
by dpflan 4y ago
Basically enough time and research has happened and cultural perception changes such that the formerly taboo science is less taboo.
604.
by dpflan 4y ago
"Say HI!" [Human Inside]
605.
by dpflan 4y ago
Your profile says all posts are composed by a LLM, are you a LLM? An autonomous agent?
606.
by dpflan 4y ago
Yeah, what skills transfer is there to a time of due dilligence that should mean something now versus being investor X in Wag’s Nth series raise because you can run digital shoulders with the big VCs and Ashton Kutcher…
607.
by dpflan 4y ago
I do agree with the environment changing and less adaptable, less useful companies will fold and faux VCs will evaporate Which I think is good. The startup scene and VC scene was clearly about hype and growth. Of the last decades lots of in
608.
by dpflan 4y ago
Im suggesting the unit tests come first, are made by the coder (maybe with assistance), and the LLM creates the feature code. And the unit tests prove the generated code works or not then you tweak. Like TDD meets LLM
609.
by dpflan 4y ago
Do you try things like make a set of unit tests and then have the LLM create methods that the tests test? It seems that there needs to be tight coupling and verification between what is generated and whether it actually accomplishes the tas
610.
by dpflan 4y ago
Ha, I was tricked too, thought the parenthetical in the title meant this was analysis by AWS, which would mean that there could be actual price and cost data...alas.
611.
by dpflan 4y ago
What does this mean? "Whereas AI in the 2010s required massive datasets and large highly specialized (machine learning) teams, AI in the 2020s requires very minimal (or no) data and minimal technical capabilities (basic software develo
612.
by dpflan 4y ago
This got me wondering about best techniques for integrating LLM code assistants into day-to-day software development, and hence Ask HN: What is your GitHub Copilot (code LLM assistant) workflow? Please share your experience here: https:&#x
613.
Ask HN: What is your GitHub Copilot (code LLM assistant) workflow?
8 points by dpflan 4y ago|4 comments
614.
by dpflan 4y ago
Hm, all right, I'm guessing that huge models as a business maybe are over until economics are figured out, but huge models as experts for knowledge distillation seems reasonable. And if you pay a super premium can you use huge model.
615.
by dpflan 4y ago
There is another variable of experience with the technology which seems to be high for Postgres, low for RabbitMQ…
616.
by dpflan 4y ago
Yeah, it seems like a more natural evolution into specialization and scale rather than a step "backwards" to PG which I suspect will be the subject of a future blogpost about replacing their pq queue with another solution...
617.
by dpflan 4y ago
Your last comment is the key, they had an issue and not the scale so a simpler approach works, but then I imagine that this company, which is a new company and growing, will have a future blogpost about switching from pg queue to something
618.
by dpflan 4y ago
Thanks for sharing! The distributed ML currently seems focused on model execution. I see another commenter's excitement about "Looking forward to NX transformations that take distributed training next level." -- which I agree
619.
by dpflan 4y ago
Sharing a comment from a similar line of discussion: """ Complexity is quadratic in sequence length. For 512 tokens it is 262K, but for 4000 tokens it becomes 16M and goes OOM on a single GPU. We need about 100K-1M tokens to
620.
by dpflan 4y ago
It’s pretty clear, the words say he was anti, then the company he helped create apparently has marketing material all about being commercialization. Unless he leaves tomorrow for the same reasons it is quite hard to disbelieve that “cash ru
621.
by dpflan 4y ago
Perhaps it’s as easy as “ethics and laws are not the same thing”. One can profit either way, but unethical profiteering may not be prevented by a law.
622.
by dpflan 4y ago
This is hilarious to me, in that the disgruntled departure just did a 180… how long until the next disgruntled spin-out for higher reasons chases the dollar too…
623.
by dpflan 4y ago
"Dario Amodei, the former VP of research at OpenAI, launched Anthropic in 2021 as a public benefit corporation, taking with him a number of OpenAI employees, including OpenAI’s former policy lead Jack Clark. Amodei split from OpenAI af
624.
by dpflan 4y ago
I thought Benioff was solving the homelessness problem... /s But yes, agreed, societal support for others has been strategically dismantled (federal and state-level) and abuses the marketing of the "American Dream" as mechani
625.
by dpflan 4y ago
This is quite interesting, the recent launches of AI API wrapper companies posted on HN draw my attention. I like the innovation and use of the latest AI technology, but at the same time, I see a wrapper around an external API. Yes there ca
626.
by dpflan 4y ago
Cool, keep building! (This isn't a feature of Microsoft 365 or Google docs yet?)
627.
by dpflan 4y ago
Seems like a great opportunity to put a LLM interface in front of this report so readers and query it...
628.
by dpflan 4y ago
Hm, April Fool's? When will Bard burn things down because its metaverse stapler hasn't been returned?
629.
by dpflan 4y ago
Do you have a full-size working prototype? Have you tried this on a real shipping container ship? And can't Maersk just copy this concept?
630.
Amazon’s Novel Packet-Switching Patent Revealed(peternorvig.medium.com)
2 points by dpflan 4y ago|1 comments

More ›