BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
SEC investigates Musk's claims re self-driving Teslas
www.huffpost.com
Elon Musk Faces SEC Investigation Over Tesla's 'Self-Driving' Claims: Report
Tesla's advice about Autopilot cars and the company's videos appear to give very different messages.
equanimous
Norse clairvoyant shapeshifter goddess
He has a big target on his back from big players..SEC investigates Musk's claims re self-driving Teslas
Elon Musk Faces SEC Investigation Over Tesla's 'Self-Driving' Claims: Report
Tesla's advice about Autopilot cars and the company's videos appear to give very different messages.
thelittleshort
Top Bloke
The autonomous agriculture revolution is going to be massive
www.linkedin.com
You’d have to think there will be plenty of room for all the major players to get a slice of the action
thestockexchange.com.au
[Video] World Economic Forum on LinkedIn: Technology is revolutionizing the way we farm. Learn more about the… | 137 comments
Technology is revolutionizing the way we farm. Learn more about the future of AI: http://ow.ly/rvjg50MC04X Verdant Robotics | 137 comments on LinkedIn
You’d have to think there will be plenty of room for all the major players to get a slice of the action
BRN Discussion Ongoing
I could well have missed this but I have just noticed today that BrainChip has a Principal Computer Scientist and he has just published a paper which He has posted to linkedin https://www.linkedin.com/in/krishnamurthy-vemuru-47383349/recent-activity/...
thestockexchange.com.au
Last edited:
Steve10
Regular
Excellent easy explanation from chatgpt. Saved me time explaining it.
Tothemoon24
Top 20
https://apple.news/AiGy_hLO5RE-MpGE83SpEIQ ‘Hey Mercedes’
Meetupsoon
Member
Good morning,
An interesting listen.
Satya Nadella: Microsoft's Products Will Soon Access Open AI Tools Like ChatGPT | WSJ
Regards
Meetupsoon
An interesting listen.
Satya Nadella: Microsoft's Products Will Soon Access Open AI Tools Like ChatGPT | WSJ
Regards
Meetupsoon
mrgds
Regular
Hi @Esq.111 and @hamilton66
Thnx for posting these two videos, @hamilton66 .............. as you said, any stock holder ......... this is a must watch !!!
I agree @Esq.111 .............. this topic most certainly should be forwarded to our management team, because if they are not aware, then they certainly should be !
As shareholders of a listed company we have the right to be looked after by the companies CEO/BOD.
So, in the interest of "being looked after" i do feel our CEO/BOD should be following this up, and also making press releases that confirm they have and will continue to be active in doing so.
We ( those who watch the daily s/p movements ) are oh so familiar with the spoofing that is conducted on a daily basis.
@Esq.111 .................. maybe the CEO/BOD could start with the "ShareIntel " that was mentioned and pay ( THE BULLDOG ) a lawyer to see if they feel further action should be undertaken.
I for one will forward this to TD and im thinking many should do the same. We have a so called thousand eyes here, and i think this definitly should be confirmed by management that they ARE AWARE and taking action, as this is systemic not only in the US.
We have had what? .............. maybe a half dozen sentences, or should i say questions ( last podcast ) from our CEO this financial year, IMO that by far is not enough shareholder engagement from our CEO.
Many will say "Oh, NDAs etc" ............... IMO theirs plenty that could and should be communicated by Sean, without going into companies persa, but just a general chat about the health of the company and the focus of efforts being undertaken.
Id personally like to see a "Monthly Newsletter" .
Again thnx @hamilton66 ..................please continue to bring the development videos of this to us on TSE
AKIDA BALLISTA
mrgds
Regular
Have forwarded both videos to TD ( tdawe@brainchip.com )
If you feel as strongly about this potentially company ruining activity, can i suggest you do the same.
AKIDA ( f*#ck the short selling ) BALLISTA
BaconLover
Founding Member
Hi @mrgds
Last AGM we heard "share price will do what share price does".
I doubt they'll use company resources to fight shorts. It's just not viable. It's time consuming, resource draining and a fight they won't win. Seriously, shorting is here to stay whether people like it or not. Manipulation is here to stay whether people like it or not.
Why bother? They should bring in sales, revenue, IP contracts and shorts will be covered and thereby share holders rewarded.
mrgds
Regular
Have you watched the 2 videos put up by @hamilton66 ?
As @esq said, ................. heres one CEO thats not going to let "THE S/P DO WHAT THE S/P WILL DO"
Now theres becoming a "CLASS ACTION " ............ awesome, get many company CEOs on task and it doesn"t cost each company much in the way of time or resources.
As i said .................. "CEOs/BODs have a obligation to their shareholders to protect the companies interest.
These types of activities can ruin companies.
Please dont stick your head in the sand,
Accepting the status quo will not change anything and im all for stamping out this corrupt activity !!!!!!!!!!!!!
AKIDA BALLISTA
BaconLover
Founding Member
I didn't say "the SP will do what SP does" quote.
A company top official did.
Good on the CHN ceo, I think there are another 2 companies fighting manipulation and defamation in my knowledge.
Just saying don't expect too much out of it. Manipulation usually takes shares from weak hands.
May be it's for the common good. Haven't done the research? Don't know what's going on with the company? Don't have the courage to ride the storm? Survival of the fittest may be for the common good.
Completely understand its not a popular opinion.
Edit: If someone has contacted Tony, please post the response if he's happy. I'd love to know company's stand on shorting. I'd say they would be concentrating on their job and not worry about markets, but wouldn't mind to hear from them.
Last edited:
OpenAI are aware of the shortcomings of ChatGPT:
ChatGPT: Optimizing Language Models for Dialogue (openai.com)
Methods
We trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup. We trained an initial model using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant. We gave the trainers access to model-written suggestions to help them compose their responses. We mixed this new dialogue dataset with the InstructGPT dataset, which we transformed into a dialogue format.
To create a reward model for reinforcement learning, we needed to collect comparison data, which consisted of two or more model responses ranked by quality. To collect this data, we took conversations that AI trainers had with the chatbot. We randomly selected a model-written message, sampled several alternative completions, and had AI trainers rank them. Using these reward models, we can fine-tune the model using Proximal Policy Optimization. We performed several iterations of this process.
ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. You can learn more about the 3.5 series here. ChatGPT and GPT 3.5 were trained on an Azure AI supercomputing infrastructure.
Limitations
ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as: (1) during RL training, there’s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.
The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.12
Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
While we’ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We’re using the Moderation API to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now. We’re eager to collect user feedback to aid our ongoing work to improve this system.
Iterative deployment
Today’s research release of ChatGPT is the latest step in OpenAI’s iterative deployment of increasingly safe and useful AI systems. Many lessons from deployment of earlier models like GPT-3 and Codex have informed the safety mitigations in place for this release, including substantial reductions in harmful and untruthful outputs achieved by the use of reinforcement learning from human feedback (RLHF).
On their CLIP page, they mention that that database took 256 GPUs 2 weeks to do the training, but I don't know if this is a reasonable comparison for GPT.
CLIP - Contrastive Language-Image Pre-training
CLIP: Connecting Text and Images (openai.com)
ChatGPT: Optimizing Language Models for Dialogue (openai.com)
Introducing ChatGPT
We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
openai.com
Methods
We trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup. We trained an initial model using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant. We gave the trainers access to model-written suggestions to help them compose their responses. We mixed this new dialogue dataset with the InstructGPT dataset, which we transformed into a dialogue format.
To create a reward model for reinforcement learning, we needed to collect comparison data, which consisted of two or more model responses ranked by quality. To collect this data, we took conversations that AI trainers had with the chatbot. We randomly selected a model-written message, sampled several alternative completions, and had AI trainers rank them. Using these reward models, we can fine-tune the model using Proximal Policy Optimization. We performed several iterations of this process.
ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. You can learn more about the 3.5 series here. ChatGPT and GPT 3.5 were trained on an Azure AI supercomputing infrastructure.
Limitations
ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as: (1) during RL training, there’s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.
The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.12
Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
While we’ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We’re using the Moderation API to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now. We’re eager to collect user feedback to aid our ongoing work to improve this system.
Iterative deployment
Today’s research release of ChatGPT is the latest step in OpenAI’s iterative deployment of increasingly safe and useful AI systems. Many lessons from deployment of earlier models like GPT-3 and Codex have informed the safety mitigations in place for this release, including substantial reductions in harmful and untruthful outputs achieved by the use of reinforcement learning from human feedback (RLHF).
On their CLIP page, they mention that that database took 256 GPUs 2 weeks to do the training, but I don't know if this is a reasonable comparison for GPT.
CLIP - Contrastive Language-Image Pre-training
CLIP: Connecting Text and Images (openai.com)
CLIP: Connecting text and images
We’re introducing a neural network called CLIP which efficiently learns visual concepts from natural language supervision. CLIP can be applied to any visual classification benchmark by simply providing the names of the visual categories to be recognized, similar to the “zero-shot” capabilities...
openai.com
Baneino
Regular
Our company has a product that cannot be found anywhere else in the world. I personally have no doubt that only through this wonderful product and hard work we will see the prices rise.
Of course I am absolutely against short selling. That does no one any favors at the moment. But in 10 years, God willing, we just laugh about it! I personally use this time just to continue to stock up.
Greetings from Germany
Of course I am absolutely against short selling. That does no one any favors at the moment. But in 10 years, God willing, we just laugh about it! I personally use this time just to continue to stock up.
Greetings from Germany
mrgds
Regular
BL ............... along with my forwarding of the information to TD, i did ask him for confirmation of when this has been forwarded to the CEO/BOD.
Of course i will post any reply i get with his permission.
AKIDA BALLISTA
Rskiff
Regular
My concern with "shorting" is it effects companies by forcing share prices down. The ability for say raising funds in the market with a lower share price (lower value of company), then shareholders will be diluted even more with the company having to offer more shares in order to raise the capital needed. Higher share price, more value in company. Lower share price, less value in company.
wilzy123
Founding Member
mrgds
Regular
As I understand it, the legal form of shorting is borrowing shares at $A per share for a fee of $B per share and hoping that the price falls to a price $C, where $(A-C) is sufficient to cover the cost of borrowing the shares ($B) plus a profit per share $D.
That is A-C > B+D.
This enables the shorter to buy the shares at $C to return the shares to the lender, thus making a profit of $D per share.
Such shorting would be legal, if unsportsmanlike.
This is a straight gamble, betting against the SP rather than investing in the company. The core business of the stock market is to foster investment in businesses. Its remit does not include a casino licence.
On the other hand, if the borrower has specific knowledge of some fact which will cause the SP to fall, and if this knowledge is not public knowledge, which could be classed as insider trading.
Furthermore, if the borrower attempts to manipulate the SP downwards by loading the sell side with borrowed shares, or by cross-trades among fellow miscreants, this would be illegal, although rarely penalized.
So, I would like to see that the borrowed shares were frozen (eg, put in escrow), out of the control of the shorter, the illegalities may be avoided, but there would be problems in determining, when the shorter requests a sale, whether this is a genuine sale. Obviously, small packets dragging the SP down on close would be a clearly criminal act. Indeed, all small packets sold with the effect of lowering the SP should be suspect.
So, in the absence of a total ban on shorting, I consider that there should be strict controls on the sale of borrowed shares.
That is A-C > B+D.
This enables the shorter to buy the shares at $C to return the shares to the lender, thus making a profit of $D per share.
Such shorting would be legal, if unsportsmanlike.
This is a straight gamble, betting against the SP rather than investing in the company. The core business of the stock market is to foster investment in businesses. Its remit does not include a casino licence.
On the other hand, if the borrower has specific knowledge of some fact which will cause the SP to fall, and if this knowledge is not public knowledge, which could be classed as insider trading.
Furthermore, if the borrower attempts to manipulate the SP downwards by loading the sell side with borrowed shares, or by cross-trades among fellow miscreants, this would be illegal, although rarely penalized.
So, I would like to see that the borrowed shares were frozen (eg, put in escrow), out of the control of the shorter, the illegalities may be avoided, but there would be problems in determining, when the shorter requests a sale, whether this is a genuine sale. Obviously, small packets dragging the SP down on close would be a clearly criminal act. Indeed, all small packets sold with the effect of lowering the SP should be suspect.
So, in the absence of a total ban on shorting, I consider that there should be strict controls on the sale of borrowed shares.
Similar threads
