- Charbel X
- Posts
- A Robot that can touch and feel like humans, OpenAI's new SearchGPT, A Startup story
A Robot that can touch and feel like humans, OpenAI's new SearchGPT, A Startup story
Shake hands with a robot and it could encode how it "felt". Warm, comfy and loose? Cold, firm and strong? A hand with Parkinson's disorder?
Good afternoon!
My day started leisurely; joined a friend for Sculptures By The Sea in Bondi, followed by a trip to little Vietnam, then been playing with an AI-powered 3D modeling program (demo coming soon 🤯) and now settling into a gentle deep dive into the world of design, tech and product.
Vatican goes anime, OpenAI is about to face off with Perplexity and Google, Meta making sensitive robots, a community-sourced phone design and so much more.
As always, lots to cover today; let’s get into it.
Good Things Come To Those Who Make.
Yours in wonder,
Charbel
Founder of Velvet Onion, Faster Zebra and more to come …
Today’s Highlights
AI: OpenAI’s SearchGPT is here: A Face-off against Perplexity and of course, Google
Design: Nothing’s new Community-Designed Phone: Featuring unique Glow-in-the-Dark design
Science & Tech: Meta’s Sensory Robot Hand: Can touch & feel like humans
Business Story: A Bootstrapped AI Startup hit $700,000 in revenue in just 3 months: Insights from the young college graduate who co-founded it
Founding: What your customers need > What your customers are: Look for cause over characteristics
Product: Product Roadmaps aren’t always the best option
Today’s AI image: AI can touch and feel
Quote for the day: Life Lesson from the Greatest Genius
AI
OpenAI’s SearchGPT is here: A Face-off against Perplexity and of course, Google
ChatGPT levelled up. Now as a user, you can expect more detailed, accurate and up-to-date answers to your chat queries because it can crawl through entire web to source your answers besides the data stocks it was pre-trained on.
Everything you need to know
The fine-tuned model will provide up-to-date information along with citations and references to relevant sources.
There is an aggregate of outsourced search providers (as Bing) and publishers (News Corp, Financial Times and more)
Not lagging behind Perplexity in visual interfaces, OpenAI is also adding such designs to this search for genre-specific prompts; weather, sports, stocks and maps.
It is presently available to Plus and Team users. Next week for Edu and Enterprise and for free users, it will roll out in a couple months.
Why is this a big deal?
Every top player in the AI industry is currently all about integrating web search to Generative AI. The advancement won’t crush the traditional search totally (not in near future) , even when it seems to be an apparent outcome. In fact, what will happen is the AI search facility will act as user’s “crawl” assistant.
How we normally use traditional search: Manually modifying and adjusting keywords in search prompts to get needed search results (hopefully accurate) and then rummaging through a hundred web pages to curate a required combination of info package.
How will it look like with AI web search: The entire process above will be delegated to the AI model.
Also in AI
Design
Nothing’s new Community-Designed Phone: Featuring unique Glow-in-the-Dark design
Nothing held a contest where it asked its community to come up with a smartphone design purely out of their imagination. About 900 ideas were proposed by the audience with the look and the marketing of the phones all customized by them.
These entries eventually converted into a marketised company-community collaboration in the form of a new phone launch.
A Peek into Nothing’s 2A Plus Community Edition Phone
The back of the phone will emit a cozy green glow in the dark for hours.
This is a result of the green phosphorescent substance inserted in the back.
According to Nothing, the mechanism of glow-in-the-dark will easily be charged through daylight.
You can purchase the phone from Nothing’s website after 12th of November. Only if the limited stock of 1000 units won’t already be sold out.
Also in Design
Science & Tech
Meta’s Sensory Robot Hand: Can touch & feel like humans
Meta is working on a highly specialised touch-perceiving technology in partnership with Gelsight, a sensor-manufacturing firm and Wonik, a robotics company.
Project with Gelsight: A tactile fingertip
In its collab with Meta, Gelsight will commercialise a tactile sensing system called Digit 360 or “a tactile fingertip” as described by Meta. Digit 360 will have two main technological aspects:
Sensing: About 18 sensory features are claimed to be integrated to sense and pick up a variety of signals from the stimuli in the environment.
Like vibrations, heat and even odour.
Digitising: An in-built AI chip will digitally process the captured signals.
Project with Wonik: the new Allegro Hand
The Allegro Hand contains tactile sensors similar to Digit 360. This new generation of Wonik’s Allegro Hand being created in collaboration with Meta will be connected to a host computer to transcribe every bit of data produced by the hand.
These devices are tailored to usage for deep scientific research and lab work. Meaning it’s not for consumers.
Digit 360 will be available for purchase by scientists and researchers next year.
The Allegro Hand will also be available by the beginning of next year.
Also in Science & Tech
Business story
A Bootstrapped AI Startup hit $700,000 in revenue in just 3 months: Insights from the young college graduate who co-founded it
Alec Nguyen was a struggling immigrant student in the USA surviving through college working multiple jobs - mainly student tutoring of subjects like economics and data science and as research assistant for his professors.
He picked a common pain point in his domain of work and made a million-dollar business out of it. Here’s some insights from his business journey:
Afforai: His AI startup
The field of Academia (Alec’s niche) is highly partitioned with respect to views on usage of AI. So if Alec was to create an AI-based solution for productive research and study, it was supposed to be ethically justified and morally correct.
Based on this, he and his co-founder launched Afforai AI, a tool that helps researchers, PHDs, post grads and students in general to effectively add references and citations to their work while they write.
Their product saw a seamless adoption success and the startup touched $700,000 in revenue in just three months.
Nailing the product-market-fit: How they catered to the bias & accuracy obsessed market of academics
Alec explains how they solved this problem in two parts-
They limited the source data for AI to only the documents uploaded by the user. Hence, every AI-generated response was highly based on the user-provided data leaving less room for missed context.
They built their own tech that finds where the AI got its answers from and then adds likewise citations and references; page, paragraph and sentence.
His thoughts on distribution in Tech
According to Alec, both distribution and product go hand-in-hand. Quality of each should align with the other.
For product, he suggests that an entrepreneur can either “launch first then reiterate” or “wait for the perfect product then launch” .
Founding
What your customers need > What your customers are: Look for cause over characteristics
Organisations often tend to set shallow traits-based criteria for their customer profiles.
“Pitch to every company whose LinkedIn page mentions ‘50-100’ employees”
“Go crazy with cold email bombarding on a company’s email address if you spot the word ‘B2B’ anywhere in their company profile”
“Approach every consumer who is above 50”
Yes, you still somehow reach the customers whose demands align with your product. However, don’t you think the process can be 10x more efficient if you approach a smaller circle that only has those customers?
You can achieve that smaller circle by finding the root cause that’s making your customers buy your product. It involves cracking open their heads (not really!) and figuring out their motivation behind buying. In most cases, it is an unfulfilled need or broadly, an unsolved problem.
Once you have that key, you can infinitely leverage it to unlock doors to rooms full of high-stake-buyers. All you have to do is approach the sales message in a fine-tuned way than before.
So decide for yourself - which one’s more productive after all, 2 conversions out of 25 sales calls OR 1 conversion out of 5 sales calls?
Also in Founding
Product
Product roadmaps aren’t always the best option
Product roadmaps trace a product’s probable evolution journey by defining the vision, direction, target and priorities. These roadmaps don’t help much when-
1.Longer-term forecasting is not possible- When the environment is too dynamic to forecast circumstances in the distant future, it is better to set short-term goals and act according to the environment's stimuli.
2.You don’t have a validated product strategy- When your product strategy is wrapped in significant risks and contingencies, a product roadmap is not a practical approach.
3.Your stakeholders consider it an apparent outcome- When the stakeholders hold the conviction that the product roadmap is exactly what must take place in the future, you should avoid making one.
Today’s AI Image
AI can touch and feel
Source - Bing Image Creator
Quote of the Day
Life Lesson from the Greatest Genius
“Out from clutter, find simplicity. From discord, find harmony. In the middle of difficulty lies Opportunity."
Albert Einstein
What we’re working on
Velvet Onion & Friends We’ve been building a complex, enterprise wide, data-rich, AI-enabled product for our friends at IAG. As their product development partners, our process with their great team has been faster, leaner and more cost-effective for the company. | We’re redefining the way students learn skills in Strategy, Business, Product Development, Design, RevOps, Marketing and Pitching. We’re soon launching our flagship program where students learn everything from strategy to pre-launch. The winners get their idea built by Velvet Onion. |