research-lab-ed-hal-matias-az1-ai-blake-proposal-review-2026-01-13-09-57-est-notes-by-gemini-1
Jan 13, 2026
Ed/Hal/Matias AZ1.AI - Blake proposal review
Invited Edward Gargano Matias Meirelles Van Vliet Hal Casteel
Attachments Ed/Hal/Matias AZ1.AI - Blake proposal review
Meeting records Transcript
Summary
Hal Casteel and Edward Gargano discussed Hal Casteel's upcoming deep dive with mentors, which Edward Gargano offered to advise on, as well as the importance of human expertise in strategy despite Kodate Tech's AI-first approach. Matias Meirelles Van Vliet provided an update on the partnership agreement, detailing a three-stage plan to onboard Will as a founding advisor with vesting equity to mitigate risk related to Will's full-time employment at Grail. The participants also discussed the equity split rationale, the finalization of the cap table and IP assignment documentation, and the use of Git for document management and version control. They also addressed Blake's compensation proposal, with Edward Gargano advising against the proposed 10% equity and 20% renewal commissions, suggesting an earning structure with 1% equity (up to 3-4% based on milestones) and a 10% commission rate on initial contracts tied to revenue recognition.
Details
Notes Length: Standard
-
Founder Location and Upcoming Deep Dive Hal Casteel checked in with Edward Gargano about their daughter and confirmed their move back to Florida. Hal Casteel mentioned they are uncertain about returning to Boca Raton immediately, while Edward Gargano anticipated the next step would be a "deep dive" with three or four mentors to assess their current stage and accomplishments, which Hal Casteel confirmed is a panel-style meeting (00:00:00). Edward Gargano was optimistic about Hal Casteel clearing this deep dive gate, and offered to sit in on the discussion as an advisor, potentially providing materials about the types of questions to expect (00:02:07).
-
Partnership Agreement and Will's Role Matias Meirelles Van Vliet updated Edward Gargano on the partnership agreement, noting that Will's full-time employment at Grail presented a potential risk due to their comprehensive employee agreement claiming ownership of creations (00:02:50). Matias Meirelles Van Vliet utilized four AI agents to analyze Will's compensation package and employee agreement, resulting in a three-stage path forward. The solution involves initially onboarding Will as a "founding advisor" with vesting equity, avoiding the title of co-founder in formal documents and slides, and then transitioning Will to full-time status with a bonus and a larger equity package once the company generates enough revenue, acknowledging that Will will not receive their current salary of $450,000 (00:03:57).
-
Equity Split and Commitment Matias Meirelles Van Vliet detailed the rationale for the equity structure, which rewards the level of commitment to the company, noting that both Matias Meirelles Van Vliet and Hal Casteel are full-time and investing their savings, while Will is contributing a minimal amount of time advisory capacity without having contributed to the Intellectual Property (IP) (00:05:28) (00:21:24). Hal Casteel emphasized this structure as a major milestone, as the previous imbalance was a potential liability (00:06:27). Hal Casteel sought Edward Gargano's review of the analysis within the next week to surface any issues before sending it to Andre for final legal packaging (00:07:10).
-
IP Assignment and Cap Table Matias Meirelles Van Vliet mentioned building the IP assignment documentation and that Hal Casteel was signing the IP to the company. Matias Meirelles Van Vliet also highlighted the need to finalize the cap table and the co-founders' partnership agreement, specifically whether Will will be included in the latter given their founding advisor capacity. Edward Gargano deferred these items to Andre, the attorney, for legal sorting (00:08:07).
-
Document Management with Kodate Tech and Git Hal Casteel explained Kodate Tech's operational structure, which uses Git as a document management and version control system because LLMs treat code and documents as language that can be tokenized and mapped mathematically (00:08:56). Git allows for a complete audit trail of all activities, treating Kodate Tech records as the system of truth, which can be published to Google Drive if needed (00:11:01). Hal Casteel instructed Matias Meirelles Van Vliet to show Edward Gargano the GitHub repository structure and mentioned that the lawyers would dive into these records for an inventory of created assets (00:12:13).
-
Venture Foundation Blueprint Matias Meirelles Van Vliet described a major project called the "Venture Foundation," which aims to develop an operational blueprint for A1/Kodate Tech and a generalized blueprint for any business to use for venture creation, addressing all non-coding processes like registration, filing for EIN, signing agreements, and opening a bank account (00:13:36). This process breaks down action items into single, reproducible steps, creating long-term value by documenting and sharing tribal knowledge with customers (00:16:04).
-
Will's Equity and Risk Assessment Matias Meirelles Van Vliet reviewed the analysis of Will's employee agreement, concluding that Grail is unlikely to claim Kodate Tech IP due to the differing industries and California labor law. The analysis suggested a founding advisor equity of 0.5% to 2%, but Matias Meirelles Van Vliet and Hal Casteel felt it should be higher, though under 10% to limit risk, which Will agreed to (00:17:59). Hal Casteel emphasized the importance of Will being comfortable with this structure to resolve an issue that can cause businesses to fail (00:19:07). Edward Gargano affirmed that the process is on the right track and that the proposed percentages are within the acceptable range (00:19:58).
-
Blake's Proposal and Board Advisory Role Hal Casteel shifted the discussion to Blake's proposal, which Matias Meirelles Van Vliet deemed unfair, proposing 10% equity (a co-founder level), a retainer, monthly payments, and success fees, lacking structured accountability (00:22:00). Edward Gargano viewed the offer as an opening "salvo" and suggested that compensation should be based on an earning structure with no cash upfront, emphasizing the need to preserve cash (00:24:14) (00:28:19). Will was strongly against the proposal due to its perceived greed (00:25:58).
-
Blake's Compensation Structure Hal Casteel and Edward Gargano discussed an earning structure for Blake that would be operational-focused, aiming for a fractional role to generate revenue (00:29:54). Edward Gargano suggested a "go to market strategic leadership role" rather than a Chief Revenue Officer (CRO) title until traction and board approval (00:31:01). Edward Gargano advised offering Blake 1% equity, earned over three to four years with a six-month cliff, and additional equity based on milestones, with a maximum of 3-4% total equity, confirming that 10% is a deal breaker (00:31:52).
-
Commission Structure and Payouts The team discussed commission rates; Blake had proposed 20% commission on initial contracts and 20% on renewals (00:33:42). Edward Gargano strongly advised against renewal commissions, stating this would drain profits. Edward Gargano suggested a 10% commission rate and stressed that payouts for contracts like a three-year, million-dollar deal would be monthly, tied to when revenue is recognized as services are delivered, and would stop if the deal is lost (00:34:24). Hal Casteel confirmed that they will use the notes from the call and Edward Gargano's email to define the company's offer to Blake (00:36:06).
-
AI Use in Decision-Making Hal Casteel asked Edward Gargano about their process, confirming that Edward Gargano uses Generative AI (Gen AI) tools, like Chat GPT, for research, but bases the vast majority (85%) of their thinking on their own experience and knowledge, as Gen AI can be inaccurate (00:36:45). Hal Casteel stressed that while the company is AI-first and heavily dependent on AI, all decisions, especially those concerning strategy and long-term perspectives, require human thought, experience, and the final human sign-off, emphasizing that this culture of accountability is being established now (00:38:33) (00:41:34).
-
Leveraging AI for Experience and Decision-Making Edward Gargano stated that AI allows them to save time on research and reinforces their own thinking, bringing historical knowledge to bear (00:39:37). Hal Casteel described Kodate Tech's advanced process, which includes running academic research and using AI to build new skills, personas (agents), and workflows (00:42:36). Hal Casteel explained the use of a "mixture of experts" for analysis and documentation, followed by a separate group of expert "judges" to analyze the quality of the advice, which helps minimize bias (00:43:45).
-
Next Steps and Administrative Documents Hal Casteel reiterated the goal of keeping documents like the cap table and partnership agreement "flat and simple" for efficiency, noting Andre, the lawyer, is currently very busy. Edward Gargano agreed to look through their files for cap table and partnership agreement templates and to send Hal Casteel a script for the follow-up negotiation with Blake (00:36:06) (00:46:50).
Suggested next steps
- Edward Gargano will work with Matias Meirelles Van Vliet over the next week to understand Will's founding advisor role and flag any issues, and will provide advice after Will gives his feedback on the proposed structure.
- Hal Casteel will nail down and sign an agreement with Will in no more than a week.
- Edward Gargano will look for cap table templates and a partnership agreement template from his old companies for the team.
- Edward Gargano will send Hal Casteel a script for responding to Blake's proposal regarding the go-to-market strategic leadership role.
- Hal Casteel and Matias Meirelles Van Vliet will reach out to Jennifer to confirm that the upcoming events are virtual.
You should review Gemini's notes to make sure they're accurate. Get tips and learn how Gemini takes notes
Please provide feedback about using Gemini to take notes in a short survey.