HN user
That’s definition #2 with #1 being:
a man who presents a child at baptism and promises to take responsibility for their religious education.
So, the original definition is a gendered and religious term.
Sounds silly to me. Especially in this context.
Sorry, who’s “they” here? The incumbents OpenAI / DeepMind / Anthropic?
Can we stop labeling prominent AI researchers as “AI Godfather”? It’s so silly and barely truthful.
The concept of AI has been around since Turing and if anyone deserves a title like “Father of AI” it’s him.
LeCun is Chief AI Scientist at Meta. They can just leave it at that.
Licensing can definitely turn into regulatory capture if it expands enough. It’s effectively a barrier to entry defined by the incumbent.
The Design of Everyday Things, by Don Norman is a good place to start.
https://www.amazon.com/Design-Everyday-Things-Revised-Expand...
Glad to hear we’re not the only ones having an emergency LangChain hackathon.
So far what we’ve been able to build feels brittle. But, LLMs are fun to play with though
Why does Bing AI sign every message with an emoji?
ChatGPT doesn’t do it and it comes off so strange.
How well GPTZero can detect ChatGPT generated text by measuring perplexity (how random the words choices are) and burstiness (how diverse the sentence structure is) shows that whatever algorithm our brain uses has stronger creative capabilities than this LLM.
GPT-3.5 isn’t a great writer like AlphaGo is a great go player. Maybe one day AI will generate better scripts and novels than humans, but not this model.
Medium-quality writing is ok for informative content though, but it’s problematic when the model doesn’t know fact from fiction. That’s the important complaint.
Is it dangerous? Maybe.
But is it useful? Not if it’s wrong too often.
You’re right that this tech should be taken seriously, but so should the hallucination problems. These problems can be solved. And maybe they should be solved before anyone trusts it with serious questions.
What LLMs does LangChain support?
Btw I asked chat.langchain.dev and it said:
LangChain uses pre-trained models from Hugging Face, such as BERT, GPT-2, and XLNet. For more information, please see the Getting Started Documentation[0].
That links to a 404, but I did find the correct link[1]. Oddly that doc only mentions an OpenAI API wrapper. I couldn’t find anything about the other models from huggingface.
Does LangChain have any tooling around fine tuning pre-trained LLMs like GPTNeoX[2]?
[0]https://langchain.readthedocs.io/en/latest/getting_started.h...
[1]https://langchain.readthedocs.io/en/latest/getting_started/g...
We're predisposed to seeing meaning, and perhaps "intelligence", everywhere.
I’m guilty of this with my dog. I can’t help it with her head tilts and deep stares! Her inner monologue is probably less sophisticated than I like to think it is.
Yes on all three points! Especially search. Notion’s search is so painfully slow. Whenever I try to link a row to another row the loader sounds for 2+ secs, even if I’m searching a table with <10 rows.
The first use case in the demo is generating a blog post, so I’d assume that they’re trying to win the “cynical marketers wanting to use GPT-3 to churn out keyword nonsense to game Google Search” market. I know of some marketing teams that use notion for content management.
I wouldn’t trust the narrator here. As the CEO, Jeremy is incentivized to exaggerate the impact. It’s yet to be seen whether Bitcoin, Ethereum or Doge will be considered securities. Neither Bitcoin or Ethereum seem to pass the Howey test [0], but who knows.
One significant case the LBRY ruling could impact is the SEC’s suit against Ripple Labs and two of its executives, who have been charged with selling $1.3 billion in unregistered securities. Much like LBRY, Ripple Labs’ defense has hinged on its claim that its native token XRP is not a security.
The precedents being set with these cases will have an enormous ripple effect on crypto and web3 companies.
Revenue/employee at Twitter is about $600k while it’s over $1.2M at Meta and Google [0]. Elon is winding up for a layoff with comments like this. Not sure if the accuracy of his comment is relevant.
[0] https://twitter.com/RyanReeves_/status/1423013707190206464?l...
It could be a bear market rally, which would not be too unusual [0]. Jamie Dimon (JPMorgan) says the market could drop another 20% before we hit bottom [1]. Then there’s OPEC threatening to cut supply. So, there’s certainly a chance of more pain to come in the stock market, but it’s impossible to predict a bottom.
I do agree with your overall sentiment though. After this economic storm passes there will likely be good times ahead for the US. There are lots of great companies selling at a discount right now and there will be more in the near future. It’s a great time for long-term investors to buy. (Apple is looking great!)
With that being said, this may represent a permanent change in the tech industry. Especially for startups and private, VC-backed companies. Last year we saw private SaaS companies raise at 100x ARR. Those multiples have dropped back down to 8-10x and may never reach similar heights again.
This could even be the end of the Unicorn era. We’ve had near 0% interest rates since 2008. This was the primary driver of high tech valuations and the proliferation of Unicorns. Now with rates closer to normal, profitability will be expected sooner from tech companies and startups.
In fact, Bessemer is trying to retire the term “Unicorn,” which is valuation-based, and replace it with “Centaur,” which is ARR-based (ARR>=$100M) [2]. It hasn’t caught on yet, but I’m a fan of the punnier spelling (CentARR).
[0] https://www.marketwatch.com/story/heres-how-big-of-a-bear-ma...
[1] https://www.barrons.com/articles/jp-morgan-jamie-dimon-51665...
D-Wave is not a universal quantum computer. It’s a quantum annealer [0].
[0] https://www.quora.com/Why-is-D-Waves-quantum-computer-not-co...
dang, a personal attack…
I’ll leave. Peace
I started to become concerned about Elon when I read his 2017 interview with Rolling Stone[0]:
I explain that needing someone so badly that you feel like nothing without them is textbook codependence. Musk disagrees. Strongly. “It’s not true,” he replies petulantly. “I will never be happy without having someone. Going to sleep alone kills me.” He hesitates, shakes his head, falters…
He’s been in desperate need of therapy for a long time. It’s sad in some ways. He is one of the most brilliant men alive, but he would be 10x happier and more effective if he worked on himself as hard as he works his companies.
[0] https://www.rollingstone.com/culture/culture-features/elon-m...
The other night I watched the Crypto Startup School video series by a16z on YouTube [0]. It includes presentations from smart people saying that blockchain/ethereum is the computing platform of the future.
One issue with ETH/smart-contracts that is not addressed in the series is how to ship patches to a smart contract. Let’s say there’s a vulnerability in my contract that accidentally leaks ETH.
How do I ship a patch to fix the contract? Could my buggy contract be taken advantage of indefinitely? Maybe I could move all my ETH to a new wallet and leave the bad contract? How is this problem handled today?
[0] https://youtube.com/playlist?list=PLK9Lwn4_TfLS3I9huJjd-k_Fe...
Medical data mining companies have made a business of scraping the details of consumers’ daily lives into medical dossiers that, if combined with MarketScan’s de-identified information, could be used to re-identify the individuals within its databases.
De-identification is unreliable. If you have enough context, then the patients can be re-identified.
I’m just noticing that my comment went from 5 points to 0 after your reply. I’m sorry if it was upsetting. Of course I know you are not the author of the benchmarks repo.
I made that comment in good faith pointing out that the author was not saying WASM is a catch all performance booster.
But I could get on board with your claim here that “almost always” is disingenuous (which wasn’t what you were saying in the previous comment).
The author isn’t saying it’s a “cure all performance solution.” The quote you copied uses cautious language like “almost always” and “results do vary.”
The author also cites real world apps that have switched to WASM and seen big performance gains (Figma and 1Password), which is much more compelling than the benchmarks you shared.
Levenstein distance with WASM is 40% faster in the benchmarks on the linked medium article.
WASM (145,086 ops/sec)
JavaScript (102,775 ops/sec)
https://medium.com/netscape/javascript-c-modern-ways-to-use-...
I believe this is more or less Christensen’s argument in “The Innovator’s Dilemma” [0]. Google is now the incumbent beholden to shareholders.
[0] https://en.m.wikipedia.org/wiki/The_Innovator%27s_Dilemma
Web3 isn’t going to be about crypto or the metaverse. It will be all about WASM.
I’m confused by the term “decentralized internet.” The internet has always been decentralized. When the internet was created it’s primary use case was for decentralized communication.
A better term might be “blockchain internet”?
people absorbing a technology as an element of personal identity
This is true for all modestly popular technologies today. It’s fueled by social media culture. We are living in a world where people get wrapped up in their individual identity because it’s the only way to get any attention. This is a simply how social capital is distributed on the internet.
If you want your programming language, new tech, new video game, or new thing to be popular, then you need to create hype with influencers that wrap their identity up in your new thing and preach it’s endless benefits with cult-like fervor.
In fact, it’s impossible for a new language to survive unless they play this game. In other words, the hype we all hate isn’t rust’s fault, it’s our fault.
This is great for blogging, whistle blowing, and other things.
Could this support user authentication for uploading content? Maybe with something like metamask?
Using this tech to make a medium/substack/twitter-style site would lower the barrier to entry for non-technical journalists.
Congrats to the Django team! Adding support for a Redis cache instead of only memcached is awesome.