Oras

34 ms·

541.
by Oras 2y ago
I found the article misleading. Microsoft has a 49% stake in OpenAI and exclusivity in using their models. They have also invested in Mistral AI, so they are not "outsourcing" their AI, as the article suggests.
542.
Microsoft is outsourcing its best AI, tech CEO says – good news for Google(cnbc.com)
1 points by Oras 2y ago|2 comments
543.
by Oras 2y ago
Are you imagining something similar to QuidCo [0]? https://www.quidco.com/
544.
by Oras 2y ago
I think Indeed will have a huge legal risk selling personal data, but it could be scammers pretending to hire just to collect resumes, and of course the personal data of the applicant.
545.
by Oras 2y ago
The article made me question Google search even more. « we tested the feature extensively before launch. This included robust red-teaming efforts, evaluations with samples of typical user queries and tests on a proportion of search traffic
546.
by Oras 2y ago
I used to dream being in an exam that didn’t prepare to, it kept happening even 4 years after graduation. I would wake up sweating and thirsty. The interesting bit is I never cared about exams in the uni, and had no issues being unprepared.
547.
by Oras 2y ago
This is well excited. I don’t know what’s the use case, is it just create a graph and download image? I’ve used a similar service before to generate graphs on the fly in order to use them inside PDF. It made prototyping quick and easy
548.
by Oras 2y ago
I would ask first: what’s the business value from implementing it? Is it going to speed up time to market for new features? How many features do you ship per year? What’s the ROI. Business do not give a monkey about tech or processes unless
549.
by Oras 2y ago
This wouldn’t work unless you are applying to a created-in-weekend service. Proper ATSs parse resumes, extract skills, work history, … etc and they did that way before OpenAI existed
550.
by Oras 2y ago
This is brilliant, I like the execution, design, and clarity of navigation. I noticed you’re using https://financialmodelingprep.com/ api. Did you face any challenges with data quality?
551.
by Oras 2y ago
From the earnings PDF, reference 2 > Due to GenAI and in-house built developer productivity tools, we have increased the output of our software developers by around 70% year-over-year. In parallel, we have been carrying out a large-scal
552.
by Oras 2y ago
I've been using a prompt management service (cloud) for 2 months and am pretty happy that I can define prompts and check quality outside the code, and it helps me to do the routing manually. Having benchmarks (I assume this is ongoing?
553.
by Oras 2y ago
It is certainly a problem but you said you value privacy, so I think it’s a hard sell for people to give access of their email (which mostly have sensitive information) to a SaaS product that has just started. Btw, there is an open source a
554.
by Oras 2y ago
Have you tried AWS textract for table extraction then LLM to format the data?
555.
by Oras 2y ago
Depends on how you define “make it” But usually the number is a lot lower.
556.
by Oras 2y ago
Market your service if you're selling a service.
557.
by Oras 2y ago
Is there a way to export the created tutorial into markdown?
558.
by Oras 2y ago
For the inference part, you can dockerise your model and use https://banana.dev for serverless GPU. They have examples on github on how to deploy and I’ve done it last year and was pretty straightforward.
559.
by Oras 2y ago
potato <=> potato (sounds on)
560.
by Oras 2y ago
Oculus still requires a Facebook login; don't you see the conflict?
561.
by Oras 2y ago
I like the categorisation, where I can see everything on one page rather than navigate multiple pages. Suggestion: Collapsable links at the left side. For example, Content Generation and knowledge exploration. I want to click, collapse&#x2F
562.
by Oras 2y ago
That modem sound!
563.
by Oras 3y ago
That's impressive. I asked to summarise an article in 5 bullet points, and the output was 812.81 T/s on Llama 3 8B.
564.
by Oras 3y ago
That’s it really, mention someone with higher reach to get more visibility
565.
by Oras 3y ago
It would be impossible to do E2E encryption unless the transcription happens on the device (I assume you mean the wearable). Even with that, you will only access the transcript and still need an AI model to get meaningful info. This device
566.
by Oras 3y ago
Not yet but planning to give it a try and compare with textract.
567.
by Oras 3y ago
I did, none of the open source parser worked well with tables. I had the following issues: - missing cells. - partial identification for number (ex: £43.54, the parser would pick it up as £43). What I did to compare is drawing lines around
568.
by Oras 3y ago
Amazon Textract, to get tables, format them with Python as csv then send to your preferred AI model.
569.
by Oras 3y ago
+1 for Claude Opus, it had been my go to for the last 3 weeks compared to GPT4. The generated texts are much better than GPT4 when it comes to follow the prompt. I also tried the API for some financial analysis of large tables, the respons
570.
by Oras 3y ago
I haven't tested the link, but I'm familiar with similar products. You don't seem to know what exactly goes behind screening CVs. To get structured data, it's not just about parsing the file (PDF, Docs) with an open sour

More ›