2,873 karma · joined November 7, 2012
I wouldn't hesitate to share the evidence with a journalist with a publication backing them though as that would be a more appropriate venue.
We do search for recorded speech using deep-learning. We find keywords/phrases and predict what's in videos, podcasts, phone calls—anything with recorded speech, really. Founded by two particle-physicists, we’re a highly-technical team working on hard-problems (building ML models for sales, anti-fraud and understanding human languages via sound) for big-enterprise (as well as some more playful tools like Hoogley for Youtube and Podenvy for Podcasts).
Read about us in TechCrunch last week @ http://tinyurl.com/deepgramtc Read our post about How to get a job in Deep learning @ http://tinyurl.com/deepgram2
Role Descriptions: Web Engineer We need help building-out the speech search API and front ends. Leans toward backend but creating good interfaces is needed. Creativity, high energy, motivation, and experience building APIs and complicated web apps is a must. Experience working with Tornado, AWS, GCP, Docker and fluency in Python, C++, Javascript is a big deal too.
AI Scientist We need help building DNN models, and running it all on a HPC cluster. Creativity, high energy, motivation, and experience building DNNs in images, speech, or similar is a huge plus. Large scale computing and low level hardware experience is a great thing to have too.
Sales We’re hiring for a variety roles: SDR, A.E. and BD
To apply, goto http://careers.deepgram.com/ or ping me directly using the info in my HN profile if you have questions.
We do search for recorded speech using machine-learning. We find keywords/phrases and predict what's in videos, podcasts, phone calls—anything with recorded speech, really.
Founded by two particle-physicists, we’re a highly-technical team working on hard-problems (building ML models for sales, anti-fraud and understanding human languages via sound) for big-enterprise (as well as some more playful tools like Hoogley for Youtube and Podenvy for Podcasts).
Role Descriptions:
Web Engineer We need help building-out the speech search API and front ends. Leans toward backend but creating good interfaces is needed. Creativity, high energy, motivation, and experience building APIs and complicated web apps is a must. Experience working with Tornado, AWS, GCP, Docker and fluency in Python, C++, Javascript is a big deal too.
AI Scientist We need help building DNN models, and running it all on a HPC cluster. Creativity, high energy, motivation, and experience building DNNs in images, speech, or similar is a huge plus. Large scale computing and low level hardware experience is a great thing to have too.
Sales We’re hiring for a variety roles: SDR, A.E. and BD
To apply, contact careers AT deepgram dot com or ping me directly using the info in my HN profile.
Bhatia had gained access to nearly everything about a user, and in a further email to another employee, he added that “I can turn any non paying user into a paying user, vice nversa, compose messages between users, check unread stats, etc."
Biderman wanted to take advantage. “Holy moly..I would take the emails...” he replied.
But Bhatia wasn't keen. “can't do it.. want to be able to look my son in the eye one day.” Bhatia did, however, demonstrate to Biderman how to complete the process, and sent a .txt file apparently containing a wealth of information on a Nerve user. The file included an email address, seemingly hashed password, and plenty of other data.
"Use Real Court Evidence to Solve the Murder of Hae Min Lee Access Courtroom Audio, Build Time-lines, Maps & Suspect Bios. Team up with fans of the Serial Podcast from around the world to help solve a murder.
Every Court Document and Recording We're Publishing Every Piece of Publicly-Available Evidence from Adnan's Trial and building powerful annotation tools"
Ethical quandaries aside (of which there are several) for a moment, I think this strategy speaks poorly of the startup. What type of signal does it send to prospective investors that you feel it's necessary to pull illegal stunts in order to gain attention for your round? A quality YC startup shouldn't need to go to such lengths to raise a round if the substance of your project/team is of sufficient quality as to warrant investment.
Ethical quandaries aside (of which there are several) for a moment, I think this strategy speaks poorly of the startup. What type of signal does it send to prospective investors that you feel it's necessary to pull illegal stunts in order to gain attention for your round? A quality YC startup shouldn't need to go to such lengths to raise a round if the substance of your project/team is of sufficient quality as to warrant investment.
Bille97... Billel... Billen... Billet... Billew...
Isn't it unlikely that so many alphabetically-similar accounts from the third-party site would use the same password for dropbox?
Thanks anyways though.