- Uncharted AI š¤
- Posts
- Sam Altman on Capitol Hill, Amazon, Anthropic, and Google This Week in GenAI
Sam Altman on Capitol Hill, Amazon, Anthropic, and Google This Week in GenAI
lTodayās newsletter intro will be written by ChatGPTās simulacra of Obama. Behold:
Ladies and gentlemen, boys and girls, welcome to The Week in GenAI newsletter! This is where we gather to delve into the astonishing world of generative AI.
Now, if you've got some remarkable products or intriguing tidbits that you'd like to showcase, I implore you, please drop them here, right now! We want to hear your voices, your ideas, your genius! Enthusiastic nods šš
And listen up, my friends! If someone dear to you has so kindly forwarded you this electronic missive, and you find yourself bewildered, fear not! Simply take a moment and subscribe right here, so you may revel in the delightful state of perpetual bewilderment each and every week! Chuckles š
So, my fellow seekers of knowledge, let us march forward into the vast expanse of possibilities! Onwards, towards a future filled with wonder and limitless potential! Inspiring music swells šš
Thank you President Obama for that wonderful intro. Now,
Also, really quick, Iām doing an ad giveaway this week. First person to send me a reply with the emails of 5 people they got to subscribe (Iāll check on my end) will get to run an ad copy on the next issue. Iāll email whoever won for their ad by end of day tomorrow and they get to run an ad for free to thousands of folks who like AI. Pretty sweet deal, no? And as for last weekās winner, as promised:
Coefficient-AI in Google Sheets
Click on the image to learn more about Coefficient and their cool competition, not to mention you could win the big bucks!
Now then, a fair warning: AIās Coming In Extra Toasty Today.
News Headlines
Sam Altman Testifies In Front of Congress š
OpenAIās Open Source Shenanigans š
Amazon Enters the Chat š¤
Anthropicās Model Can do What?! š±
Googleās Got The Goods š
Bodacious AI Stuff
Recipease: Anyone ever run into the problem of not being able to figure out what to make when they look into their fridge (more often than not for me). Well, Recipease is here to help you clear out your fridge like never before with AI-generated recipes created from any ingredients you can scrounge up. Delicious.
Maven: Reading about the developments in AI is all well and good, but at some point you need to start using it. Maven wants to help teach you how to leverage existing AI tools to accomplish a variety of objectives, like how to prompt engineer or even start your own AI business.
āļø An AI Flyby for the Week āļø
Sam Altman Testifies In Front of Congress š
Regulating AI seems to be one of the first things our partisan government can almost singularly agree on in a long time. I feel like I must be dreaming.
Sam Altman also believes that AI needs to be regulated, recommending the creation of a new government agency that would issue licenses and establish safety standards that models must abide in order to be released to the public. Huh?! A Tech CEO advocating for government regulation? Interesting.
I have some of my own thoughts on why OpenAI might be taking this pro government regulatory stance. With the rapid rise and proliferation of Open Source AI models, OpenAI, Google, and other privately held AI companies may begin to fall behind. This advocacy for government regulation could be an attempt to slow down the open source community and allow the private sector to maintain their edge. Or OpenAI could genuinely have everyonesā best interests at heart and actually believe that government oversight is necessary for things not to get out of hand. Hard to tell with CEO Sam sometimes.
The testimony also brought up three main concerns lawmakers had about AI:
1) It can be used to create very targeted and deceptive misinformation surrounding elections which can be debilitating to our democracy. CEO Sam shared this concern heavily.
2) Copyrighting and licensing issues were also at the forefront of the discussion given the surge in the number of lawsuits over some AI models being trained on what is technically copyrighted data.
3) The effect of AI on the economy and the job market. CEO Samās take: AI will certainly take some jobs, but it will also augment dozens more and create jobs that are better than the ones that were automated.
OpenAIās Open Source Shenanigans š
OpenAI plans to release an open source model in the coming months. Details are somewhat scarce, but the model is expected to be comparable in performance to ChatGPT, but not at the level of GPT-4. This would probably mean programmatic access to a ChatGPT-level model for less than what the OpenAI ChatGPT API currently costs, which could prompt a whole new surge of application development.
This move comes amid criticisms leveled at OpenAI for originally being all for open source, but switching up around the time of GPT-3ās release. This open source model might be a reconciliatory gesture to show that OpenAI is still committed to collaborating and maintaining some level of transparency with the public. Of course, as with any story, there are always two takes.
Open sourcing their model could also help it undergo explosive growth as the open source community, again and again, has shown their willingness to improve released models (like they did most notably with Metaās Llama). After the open source community works their magic, OpenAI might simply piggyback off of the progress made and incorporate advancements into their models down the line (models which will most likely be proprietary). So, nobody really knows OpenAIās actual strategic objectives through this act, but itās safe to assume a little bit of what I mentioned is at play.
Oh, also, this probably deserves itās own section, but OpenAI announced this week it will be opening up plugins to all of their ChatGPT Plus subscribers (the number of plus subscribers is unknown). Thatās big as this means ChatGPT will be able to surf the internet which vastly improves its ability to generate up to date content on demand.
Amazon Enters the Chat š¤
Allās been quiet on the Amazon front for a couple weeks after the release of their LLM Titan and their AI fine tuning platform Bedrock (newsletter coverage from yours truly). Now, Amazon jumps back into the fray.
The news on the block is that Amazon wants to reimagine their website search and upgrade it to be more conversational in nature. That means, you guessed it, adding some generative AI to the mix. The new conversational search engine (do we still call them engines?) is supposed to also include functionality like product comparisons and personalized suggestions. However, details are sparse and a date hasnāt been set for the public launch of this feature. However, they are hiring, so if weāve got any whiplash smart conversational AI guys in the audience, nowās your chance. If you get the job, tell Andy the AI Bloke sent you.
Also, who wants to be the one to tell Amazon, that this company, Claros, kind of already beat them to the punch on building a conversation bot indexed on Amazon. Not meā¦ oh wait.
getproduct.help
#ConversationalAI
#ChatGPT
@amazon
ā One Of The AI Guys (@TheAIBloke)
7:51 AM ā¢ May 18, 2023
Anthropicās Model Can do What?! š±
You didnāt forget about little olā Anthropic (valuation sitting around 4-5 billion) right. It may be valued at ~5 times less than OpenAI, but their new version of their generative text model, Claude, can intake 3 times the amount of context tokens as GPT-4, sitting at a whopping 100,000 tokens to GPT-4ās 32,000. To put that many tokens in context, it would take an average human 5 hours to read that much text, whereas Claude can process and formulate a response in around a minute. But that means itās still not as fast as this guy who claims he can read 1000 pages in 2 seconds š. Iām just as dubious as the judges in this video, but if he really can somehow do that good for him:
With a model able to intake this much text and effectively process it so fast, there are a host of implications. Analysts who spend their days reading over dense reports can massively speedup the time it takes for them to make an assessment. This model also makes āchat with your PDF or fileā tools pretty much obsolete as the entire pdf can just be fed into a session without the need for working around token limitations. Sorry.
Googleās Got The Goods š
Well, everyone likes a good underdog story, but you got to understand the top dog is there for a reason. Watch this video to see the highlights of what Google has launched since their IO conference.
šØICYMI IN AI: Google made a ton of big announcements at its I/O keynote yesterday:
- Gmail Help Me Write
- Google Map Immersive View
- Magic Editor
- AI Generative Pics & Cinematic Wallpapers
- Bard Updates
- New Multimodal LLM PaLM 2Watch the Highlights in under 10mins!š½ļøā¤µļø http
ā āVinnie Moura (@vinniemourax)
12:50 PM ā¢ May 11, 2023
And this thread has some of the specific changes that went into place with Bard, although if you canāt tell itās a little biased.
ChatGPT is in trouble.
Google's free competitor Bard just got a MASSIVE new update.
Here's 8 things Bard can do that ChatGPT can't:
ā The AI Solopreneur (@aisolopreneur)
12:21 PM ā¢ May 13, 2023
Parting is Such Sweet Sorrow (But Please Keep Reading LOL)
Well that's a wrap on the newsletter. Before I launch into my ending monologue, please take a moment to drag this email to your primary, so that you never miss it each week! Thank you!
Let me know how I did by filling out this feedback form or just reply to the email like you would any other. If I did well and you're in the giving spirit please support the newsletter by buying me a coffee or donating to my gumroad page. Most of the money you donate to me will go back into purchasing AI products to try out and review on the newsletter. Bernie says it better:
Also if you have a product that you would like to be featured on the newsletter remember to add it here!
Lastly, if you'd like to refer a friend just tell them to visit chatgptnewsletter.com to subscribe! Or you could always just forward them the email without any subject header. I'm sure they they would love that.
Cool Things You Should Try/Buy
(Newsletter 1) ==> Consult domain experts in your browser with ExperAI
(Newsletter 2) ==> buy a talking book from Konjer
(Newsletter 3) ==> Pay for DoNotPay lol. Irony aside, it really is pretty cool.
(Newsletter 4) ==> Multion.AI - not your average burger buying extension.
(Newsletter 5) ==> Why settle for regular email when you could have Intellimail?
(Newsletter 6) ==> The AI app store where you can Cookup just about anything
(Newsletter 7) ==> A cute, fuzzy AI-powered online meeting summarizer: Otter.ai
(Newsletter 8) ==> Tired of taking hours to find the perfect online purchase? Getproduct.helpās
got you covered
(Newsletter 9) ==> Turn your Scribbles into art with ScribbleDiffusion
(Newsletter 10) ==> Ever wanted a search engine for your entire online life? Rewind.ai
(Newsletter 11) ==> Your Online Image Studio š¬: Clipdrop
(Newsletter 12) ==> Use Generative AI to create online courses on any topic you can think of. For example, you can create lesson plans about yourself before you get famous. Learn.xyz
(Newsletter 13) ==> Replace Siri with ChatGPTā HeyGPT!
(Newsletter 14) ==> Autonomous AI Agents in Your Browserā AgentGPT!
(Newsletter 15) ==> Create your own AI music with musicfy
(Newsletter 16) ==> Detect AI written content more reliably with GPTZeroās browser extension.
Reply